Wednesday 27 March 2019

Common Issues trial: the judge's comments - part 4 - the Post Office's attitude problem


This post contains extracts from paragraphs 559 - 724 of the Bates v Post Office Common Issues trial judgment. They are being broken down in this way in order to provide digestible chunks for the casual reader.

The extracts in all these posts are chosen because I thought they were of interest. They are therefore subjective and do not give the whole or complete picture of the judgment. If you want that, read the whole judgment.

This post mainly covers the Post Office's general attitude to Subpostmasters and the infamous 12:12 clause of the Subpostmaster contract:
"The Subpostmaster is responsible for all losses caused through his own negligence, carelessness or error, and also for losses of all kinds caused by his Assistants. Deficiencies due to such losses must be made good without delay."
My commentary is in bold.

On the Post Office's conduct, attitude and culture of excessive secrecy:

523… The Post Office appears, at least at times, to conduct itself as though it is answerable only to itself. The statement that it is prepared to preserve documents – as though that were a concession – and the obdurate [sic] to accept the relevance of plainly important documents, and to refuse to produce them, is extremely worrying. This would be a worrying position were it to be adopted by any litigant; the Post Office is an organisation responsible for providing a public service, which in my judgment makes it even worse. 


559. There is no doubt that the Post Office is acutely aware of the potentially damaging repercussions in terms of its reputation, if this litigation succeeds. That is entirely understandable. I have dealt with the cautions urged upon me.. and in my judgment on the pre-trial application seeking to strike out evidence I also identified that the Post Office’s own evidence for that hearing explained that it wished to avoid adverse publicity.

560. What is less understandable is the way that this approach seems to have affected the Post Office’s approach to documents....

561. ... [there is] a culture of excessive secrecy at the Post Office about the whole subject matter of this litigation. They are directly contrary to how the Post Office should be conducting itself. I do not consider that they can be a sensible or rational explanation for any of them.

On the Post Office's legal deviousness:

671. This argument [on the construction of Section 12:12] by the Post Office is, in my judgment, both circular, and is an overly intricate attempt to sow confusion and obscure the true issues in the case.

On the Post Office's attitude to branch discrepancies:

569.40 [The Post Office] sought recovery from the Claimants for apparent shortfalls. I would also add that on the evidence the Post Office did this regardless of whether disputes had been reported to the Helpline or not. This was accepted by all the Post Office witnesses, and occurred whether the SPM in question was appointed under the SPMC [Subpostmaster Contract] or the NTC [Network Transformation Contract], even though the terms of those contracts were different. It was also done regardless of any analysis of any causative fault on the part of SPMs [Subpostmasters]. It was also done when the SPM in question had been told that no action would be taken in respect of a disputed shortfall. 

On the infamous Section 12:12 of the Subpostmaster contract:

646. I consider that the correct meaning of Section 12 Clause 12 is that the SPM is responsible for losses that are caused by his own negligence, careless or error, and that it is for the Post Office to prove that any particular loss falls within that category. I consider that this is clear on the natural meaning of the words. 

649. ... it was joined with the inference adopted by the Post Office regarding the Horizon system, which is that Horizon-generated losses do not occur. The Post Office therefore argued that it is entitled to conduct itself as though shortfalls are the responsibility of SPMs, unless the SPM could prove the loss was not caused by his or her negligence, carelessness or error. 


653. In my judgment... it is the Claimant’s construction that I consider to be the correct one. It is not for a SPM to demonstrate that there was no negligence, carelessness or error on his or her part. It is for the Post Office to demonstrate that there is.

654. I consider there is a contractual burden upon the Post Office to demonstrate that the loss for which it seeks to hold a SPM responsible was caused by the negligence, carelessness or error of the SPM. 

On the impossibility of a Subpostmaster proving they were not at fault for a branch discrepancy:

655. For a SPM to demonstrate they were not at fault, if there was a loss, could be verging on nigh on impossible. Firstly, they would have to concentrate upon and analyse all of the branch records for every single transaction within the particular trading period. That would be an onerous burden for a single SPM. Secondly, those records would only be between the branch and the Post Office; SPMs have no access to data between the Post Office and its clients, and are not able to obtain it. A further factor – though far less important - is all of this would have to be done whilst the SPM was complying with their obligation to open and run the branch during working hours during the week. I am also confident, from the evidence in this trial but especially that of Mrs Stockdale, who appeared in my judgment to have done all that was humanly possible in order to work out what was happening, that a SPM simply does not have access to the type of information that would make such an onerous exercise possible even in theory. 

On why the Post Office has a clear responsibility to prove a discrepancy is the fault of a Subpostmaster (this alone could make a lot of the convictions unsafe):

656. For the Post Office, however, to do this, would be less difficult. All of the transactions of that branch could be analysed and compared with the relevant data between the Post Office and what it calls its clients, which include companies like Camelot. Secondly, that workload would be borne by the (or a) relevant department, and not by one sole individual, who would be trying to run a branch Post Office during normal business hours at the same time. Thirdly – and in contractual terms this is less important, but it is an implication of a rival construction, and hence appropriate to consider it - the Post Office is responsible for compiling evidence for prosecutions for offences of false accounting and theft by SPMs. As such, it bears the burden of proof of demonstrating to the appropriate criminal standard that conduct by SPMs is dishonest. It must therefore have the ability to demonstrate how a loss has been caused. 

665. One has only to state the Post Office’s proposition to see how far it is from commercial reality. It can only be arrived at by treating the single clause as though it were two entirely separate and distinct provisions, and by construing the second in complete isolation from the first, and by reading the second part as though it included words that are not there. 

On the Post Office's investigations department acting like a sort of comedy goon squad:

723. I agree that there is a lot to be desired from the Post Office’s behaviour as identified in the cases of the Lead Claimants. I shall give four examples only.


1. Even though the Post Office’s own case on the relevant provision in the SPMC dealing with liability for losses requires negligence or fault on the part of a SPMC, this was routinely and comprehensively ignored by the Post Office, who sent letters of demand for disputed sums in express terms as though the SPM had strict liability for losses. These letters entirely misstated the legal basis of a SPM’s liability, even where they had been appointed under the SPMC. 


2. Legal representation is not permitted by the Post Office at interviews which deal with whether a suspended SPM is to have their engagement terminated – which effectively ends that part of their livelihood. Regardless of whether this is justified or not, the specific grounds and proper particulars of why they face potential termination are not even clearly identified in advance to the SPM in question. Additionally, information directly relevant to the grounds (or at least what the Post Office is concerned about, in the absence of properly identified grounds) is not provided to the SPM either, or at least not in the case of the Lead Claimants who faced such procedures. Mr Abdulla tried at his interview to explain the situation regarding TCs [transaction corrections] and the Lottery. He was disbelieved. The documents available in the trial show that, whatever else he had done, he was telling the truth about the existence of these TCs. Neither he nor the interviewer had this information available to them at the time. 


3. I have already dealt with what happened to Mrs Stockdale after she was one of the first Claimants in the litigation. Mrs Stubbs’ evidence, which I accept, was that the Temporary SPM who replaced her was told to destroy all documentation in the branch that related to her appointment. There can never, in my judgment, be any sensible rational for such destruction of important documents, and I cannot understand why the Post Office would wish to behave in such a way. 


4. The approach of the Post Office is to brook no dissent, and it will adopt whatever measures are necessary to achieve this. An example of this is in the Modified SPMC, which in Section 15 clause 19 deals with something called an Investigation Division Interview. This Division includes investigation of potential criminal offences against the Post Office. One part deals with the presence at such an interview of a friend of the SPM. The relevant clause states: 

Modified SPMC Section 15 clause 19:
“A friend may only attend and listen to the questions and answers. He must not interrupt in any way, either by word or signal; if he does interrupt he will be required to leave at once and the interview will proceed without him. Whatever is said at the interview is to be treated as in strictest confidence. The friend may take notes of the interview but he must keep the notes in the strictest confidence. The only communication the friend is entitled to make on behalf of the person who has been questioned will be in the form of a written "in strictest confidence" statement which may be submitted by the latter, in support of any official appeal which the person questioned may desire to make in connection with the methods followed at the enquiry. No other communication about the interview is allowed (unless made by permission of the Post Office) as it might constitute a breach of the Official Secrets Acts.”
 

Other parts of Section 15 deals with the requirement for a caution and so on, but I find it somewhat unusual, and potentially oppressive, that the Post Office could seek to use the Official Secrets Acts in this way. I do not see how, in a routine case, these Acts could possibly apply in the way suggested by the Post Office in this contract. 

****************************

The remainder of the judgment (paragraphs 725 - 1122) deals with the 23 Common Issues at the heart of the first trial, and who the judge finds for, with lots of working out. I'll get something up about that soon.

More judgment-fisking:







Monday 25 March 2019

The Smoking Gun: or, "Whither Gareth?"


On a day of stunning drama at the High Court last week, we also had a series of startling admissions about bugs in the Post Office's Horizon IT system.

Before the recusal hand-grenade was lobbed into proceedings, Torstein Godeseth, Horizon Chief Architect at Fujitsu, was being cross-examined by Patrick Green, QC for the claimants.

Mr Green's cross-examination was the fruit of what looked like several months work by the claimants' legal team. They had painstakingly put together a series of conclusions based on evidence disclosed, which they then put to Mr Godeseth. It delivered, to my mind, the first concrete evidence of a smoking gun - an error generated outside a branch blamed on a Subpostmaster - and it raised the question as to how many more have slipped through unnoticed.

The financial discrepancy was mistakenly generated in 2007 by a Fujitsu engineer who was trying to replace a missing line of code in a Subpostmaster's terminal without the Subpostmaster's knowledge or permission.

During a bungled attempt to fix the glitch, the engineer put in an incorrect manual entry, causing a $1000 discrepancy in a forex transaction. It wasn't picked up by Fujitsu at the time, or rather the discrepancy was picked up, but the source of it wasn't, so it was blamed on the Subpostmaster. The incorrect manual entry was only spotted during this litigation whilst the claimants' legal team were preparing for the Horizon trial.

That's the short version. The detail below exposes the range and depth of the Horizon errors the claimants' team were able to explore.

Small beer

Let's start with a reminder of the Post Office's general position on errors in Horizon. On the first day of this Horizon trial (11 March 2019), each QC gave their opening statement to the court. Mr de Garr Robinson, for the Post Office, made reference to a joint document prepared by the two independent IT experts contracted by the claimants (Jason Coyne) and the Post Office (Dr Robert Worden) to look at the Horizon system. It is called "Table of bugs/errors/defects with acknowledged or disagreed evidence of financial impact."

There are 29 bugs listed of which 12 are agreed. As part of the table, Dr Worden estimates that the total financial impact of the agreed bugs was £165,000 and the possible financial impact of each bug on each affected Subpostmaster and would be no more than an average of £13,800. This was before any corrective action was taken to compensate the Subpostmaster and fix the bugs.

£13,800 is potentially significant to an individual Subpostmaster, but in the general scheme of things, small beer. Especially as Dr Worden calculated the likelihood of a Subpostmaster being compensated for each individual bug once it had been discovered, and concluded that the bigger the bug, the more likely it would get rectified and the Subpostmaster compensated. In Mr de Garr Robinson's words:
"In a system handling the volumes of transaction that Horizon handled, which was used by the number of branches that used it, over a period of 20 years, bugs of this number and scale can have no material impact on the overall robustness of Horizon, or putting the point more pertinently, the chances of one of these bugs affecting a given set of accounts is vanishingly small....
"... that's not to say that a claimant could not have been hit by a bug.  As I hope I have made clear to your Lordship, Horizon is not perfect. It remains a possibility, but the important point is how unlikely it is."
So there you go. The claimants believe a total of £18.7m has gone missing from their branch accounts due to Horizon errors. The Post Office believes it could be nil.

Remote access

Also in his opening statement, Mr de Garr Robinson addressed remote access, vehemently denied to the BBC's Panorama programme in 2015, then admitted to the court by the Post Office in 2017.

The issue of remote access, said Mr de Garr Robinson, was something of a red herring:
"These remote access issues are here because of concern expressed that Fujitsu... was using its administrator rights in such a way and on such a scale as to undermine the reliability of the claimants' branch accounts...
"In the real world this is...  not going to have a bearing on the robustness of the system and on the reliability of any given branch accounts in any given month.... There's a sense of unreality.... even if the claimants' case succeeds at its highest, it's never actually going to have a material impact."
Let's have a look at the evidence, shall we?

Whither Gareth?

Torstein Olav Godeseth certainly looks the part. Tall, pencil thin, with a slightly pinched expression, high cheekbones and thinning hair, he is an Oxbridge graduate who has worked on and off for both the Post Office and Fujitsu's IT departments since 1987.  He was a technical advisor at the Post Office when they, together with the Benefits Agency, procured the Horizon system. Now, he is Fujitsu's Chief Architect on the Post Office Account. He spoke in court with a a broad West Country accent, which was a surprise. There can't be many Torstein Olavs who sound like they come from somewhere outside of Bristol.

On the first day of Mr Godeseth's evidence, Patrick Green QC started the standard cross-examination practice of ripping a witness statement to shreds. In this instance Mr Godeseth had provided three witness statements, statements two and three correcting or clarifying elements of their predecessor(s).

Mr Green was interested in how much of Mr Godeseth's evidence was actually his evidence. A lot of it appeared to come from a series of conversations Mr Godeseth had had with his colleagues at Fujitsu, particularly a certain Mr Gareth Jenkins.

Gareth Jenkins was the Post Office's expert witness at the Seema Misra trial in late 2010. He was also present at a meeting between Post Office and Fujitsu (also in late 2010) where fixes to a long-standing Horizon bug (which had not been disclosed to the Misra trial or the Subpostmasters it was affecting) were being discussed.

When it came to Mr Godeseth's evidence about the Riposte system, a database used by the pre-2010 iteration of Horizon (known as "legacy" Horizon), Mr Green asked:

"it is fair to say that really the source of that information that you're giving is Mr Jenkins, isn't it, because you were working for Post Office at the time?"

Mr Godeseth agreed, saying: "I feel I had a pretty good knowledge of how Riposte works, since I needed it when I was working on Impact, but it's absolutely the case that I would get more detailed information from Gareth."

Mr Green read out a few more paragraphs Mr Godeseth had written about Riposte, focusing on his source of knowledge:

"I understand from Gareth that each message included three key pieces of information ..."
"I also understand from Gareth that messages also had an associated 'Expiry Date' ..."
"I understand from Gareth that due to the size of the Post Office network..."
"I understand from Gareth that the audit application..."
"I also understand from Gareth that once these files had been written ..."

Mr Green wondered if Mr Jenkins might be a better person to talk to about Riposte. Mr Godeseth agreed.

Mr Green moved on from Riposte to the issue of remote access to branch terminals. He quoted from Mr Godeseth's first witness statement.

"In Legacy Horizon, any transactions injected by SSC [Fujitu's Software Support Centre] would have used the computer server address as the counter position which would be a number greater than 32, so it would be clear that a transaction had been injected in this way."

It transpired that although he didn't say so in this case, Mr Godeseth had got his information from Gareth Jenkins. "So" asked Mr Green, "we could write in there "I understand from Gareth"?"

"Yes." said Mr Godeseth.

Sadly, it turns out Mr Jenkins had not given Mr Godeseth the correct information about counter position computer server addresses. It could be a number below 32, so that it wouldn't be clear it had been injected by someone from Fujitsu.

Was Mr Godeseth shocked by this? No, he wasn't shocked, that would be too strong a way of putting it.

Was Mr Godeseth aware it was a contentious point within the litigation? Yes, he was.

And was he aware that he had stated the information as fact within his witness statement without any reference to the fact he had received the information from Mr Jenkins? Mr Godeseth accepted this.

Mr Godeseth had managed to put the correct information in his second witness statement, after a discussion with a different Fujitsu colleague.

Mr Green was interested in this. Had Mr Godeseth gone back to Mr Jenkins and discussed the incorrect information? No, said Mr Godeseth, he hadn't.

Was he not interested in why he had been given duff information?

Mr Godeseth repeated his previous answer. He had not spoken again about it to Mr Jenkins.

Mr Green took Mr Godeseth back through his long association with Horizon and asked if he considered himself to be "very knowledgeable" on the subject of remote access. Mr Godeseth agreed he considered himself to be very knowledgeable. Mr Green wondered if he should have been surprised when Mr Jenkins told him that remote access of the sort they had been discussing was possible. Mr Jenkins replied:

"If you are looking to support a large system, then I think the logical conclusion is it's inevitable that you have to do this sort of thing on occasion."

Mr Green asked why this statement seemed to directly contradict his witness statement. Mr Godeseth refused to accept that it did.

Legacy Horizon ------------> Horizon Online

The second phase of Mr Godeseth's cross-examination took place during the afternoon of Wednesday 20 March. Here, Mr Green and Mr Godeseth dwelt for some time on the migration from legacy Horizon (where transaction data was stored on Horizon terminals in branch and uploaded to the servers overnight) to Horizon Online (where branch transaction data was updated to the servers in real time).

Mr Green started listing some of the problems which individual Subpostmasters told the Post Office they were having when Horizon Online was rolled out eg:
"On the 1st of March at the close of business we found that on node 5 [each Horizon terminal within a branch is known as a node] the cash was short of £1,000.  All of the figures for that day match the figures presented at the time of each transactions.  An instant saver withdrawal of £1,000 was transacted that day, but I was unable to find this transaction using the online report facility.  I feel very anxious as I believe a system error has occurred at the time of this transaction."
and
"On the 2nd of March a transaction for a cash withdrawal was completed where the system commanded a member of staff to issue the money to the customer on screen but the receipt printed for that transaction printed out a decline slip. The customer was honest enough to bring back the decline receipt a day later with the money."
Mr Godeseth agreed the migration was traumatic for some, and that when he went over to Fujitsu during the rollout Fujistu was in a state of "red alert".

Mr Green asked: "Can we infer from "red alert" that it is quite serious?"
Mr Godeseth replied: "Oh, it was very serious."
"And can you tell his Lordship why it was serious?"
"There was an issue with Oracle which was the biggest problem, but clearly there were other problems going on at the same time because it was a brand new system, but the big one was an Oracle issue which I got involved in."

Mr Godeseth couldn't remember any other specific problems with the Horizon Online rollout, but did volunteer: "there certainly were issues with the software in the early days."

Mr Green wondered why there wasn't a single reference to any of these problems in any of Mr Godeseth's witness statements. Mr Godeseth replied he was "looking to give an overall explanation as to how Horizon works."

The morning after the night before

As Mr Godeseth finished his first day of cross-examination, he was given some homework, or very possibly, hotelwork.

He was handed a KEL and a PEAK to study. A KEL is a Known Error Log generated by Horizon support level two. A PEAK is developed by Horizon support level three or four to manage, investigate and hopefully fix the known error identified by the KEL. A PEAK can list notes left about the error by helpline support, plus the discussions between the engineers at Fujitsu who are looking at it.

On his return to the witness box on the morning of Day 8 of the trial, Mr Godeseth was asked about the documents he had been asked to study.

The error described in the documents (and referred to in précis in the introduction to this blog post) occurred in 2007 at a post office branch during a foreign exchange transaction for US dollars.

This error has not been reported by the branch Subpostmaster, but has rather been picked up by the system. So far so good.

For some reason, during the forex transaction, the computer chose not to write a line of code into the Horizon message store creating a surplus of $1000 (around £484 at the time).

A fix was proposed - go into the Subpostmaster's computer and manually enter the line of code which was missing. This was done over a period of 10 minutes at 5pm on the 5 Dec 2007, and as Fujistu engineer Andy Keil adds to the PEAK: "Worth noting that the branch did not have any issues with the mismatched transactions because this was fixed before they did the roll. The branch is not aware of this and it's best that the branch is not advised."

Mr Green paused to ensure collect Mr Godeseth's understanding of the situation:
"you would fairly accept, wouldn't you, that although there are documents elsewhere that show that sometimes SPMs were advised, this shows an occasion when the SPM at this stage has not been advised, doesn't it?"
Mr Godeseth was precise: "I don't know that it shows that, but it's a fair inference from it. I think I would say that Post Office were well aware of this and I would argue that it's a Post Office decision whether or not to tell a subpostmaster."

The fix goes in. But, instead of writing the proposed:

"[Quantity]:-1, SaleValue:-484, PQty:-1,000 with, other attributes (including exchange rate) as before."

... Mr Kiel manages to update the POLFS feed for the branch with a sale value of 1,014.73 and PQty of 2,080. This has the effect of just over $2,000 being inserted in the Post Office system, generating a $1000 loss at the branch.

An update to the PEAK by Fujitsu's Anne Chambers is recorded on 14 Dec 2007:
"The counter problem which caused the first issue has been corrected by inserting a message into the messagestore, for equal but opposite values/quantities, as agreed with POL... Once the problem was corrected, there should have been no impact on the branch.  However, it has been noted that the stock unit BDC had a loss of $1,000, which was generated after the correction was made."
Don't worry, says Ms Chambers: "We have already notified Gary Blackburn at POL ... this appears to be a genuine loss at the branch, not a consequence of the problem or correction."

Fujitsu failed to spot their own engineer's mistake.

Mr Green asks Mr Godeseth if that is his reading of what happened. Mr Godeseth replied:

"Having read this PEAK in more detail overnight then yes, clearly that is what appears to have been the case."

The smoking gun. It's one of those mysterious Horizon errors we kept being told about back when I was putting together my first report on this subject in 2011. The sort of thing the Post Office told us was nonsense and then sent us legal letters to make sure we knew how serious they were about it.

This morsel was just a curtain-raiser (to mix metaphors) for the rest of the day's cross-examination.

The Callendar Square bug

Mr Green described the Callendar Square bug, which affected legacy Horizon users, and was a problem with Riposte (the database Horizon was built on, owned and operated by an American firm called Escher). It could cause discrepancies in branches and it was known internally as the "Riposte lock or unlock" problem. So far, so uncontroversial

Using PEAK and KEL reports, Mr Green noted examples of lock/unlock problems reported by Subpostmasters in 2000, 2003, 2004 and finally at the Callendar Square branch in 2005. Mr Green read out a few of the PEAK reports from those periods and posited the suggestion that "we can see that although Callendar Square experienced this bug in 2005, the genesis of it appears to go back to February 2003 and similar lock agent problems back in November 2000, is that fair?"

Mr Godeseth agreed.

With regards to one of the branches affected in 2004, the KEL report read: "Transfer in accepted Twice, thus causing a Discrepancy." The discrepancy totalled £22,290.

The "solution" was "currently with Escher for investigation."

Mr Green listed a few more reports:
"This has resulted in a loss of 3,489.69.... Phoned the PM [Subpostmaster] to explain what the problem was.  He is concerned about other transactions which he has input twice (3 Giro deposits and another cheque) because of the replication problem.  Have advised him to contact the NBSC [National Business Support Centre - one of the Horizon helplines] as this is a business issue."
"When clerk accepted transfer in it accepted it. After she had accepted it she went back into transfers and node 4 transaction for £608.13 appeared."
"PM was trying to transfer £2,490 from node 7 onto node 2.  She states that she has accepted the transfer on node 2 but the system is not showing this... On node 7 it is showing pending transfer but it is not showing on node 2... It appears on her transfer sheet as completed."
"This office had severe problems balancing on week 25, resulting in a shortage of £6,414.46...."
These are not insignificant sums and not insignificant problems.

Mr Green picked up on the latter case. The Subpostmaster is convinced the discrepancy is a Horizon error and he is getting the support of his area manager. However the Horizon helpline muddies the water:
"They told the SPMR that they would report to NBSC that they had identified and rectified the problem and that the amount could be held in the suspense account.  However, as part of the shortage relates to transfers, and no error notice will be issued, then the Suspense Account Team are not prepared to authorise the entry."
Mr Green suggested it wasn't easy for the Subpostmaster.

"No, indeed," agreed Mr Godeseth, it was "a horrible position to be in."

Mr Green continues reading from the document "No follow-up was received from [Post Office's] Service Support regarding this call."

On we went with the Callendar Square bug. Mr Green read more internal logs noting that Subpostmasters were being advised of odd workarounds:
"Discussed ongoing issues of loss due to Horizon misbalance. SPMR concerned that he has now made a fraudulent entry in that he has rolled over to the next trading period and put the loss into  suspense. He has then gone on to state that the cash has been made good, which it hasn't.  This was done on the advice of the Helpdesk."
Some Subpostmasters were being met with flat denials that there was anything wrong with the system at all:
"[Subpostmaster] said that he was having problems again with transfers.  He has contacted the Horizon helpdesk who have subsequently come back to him to say that there is no system problem and that he should contact NBSC."
We finally get to an internal log written about the bug by Anne Chambers from Fujitsu in Feb 2006:
"Haven't looked at the recent evidence, but I know in the past this site had hit this Riposte lock problem 2 or 3 times within a few weeks.  This problem has been around for years and affects a number of sites most weeks, and finally Escher say they have done something about it. I am interested in whether they really have fixed it which is why I left the call open - to remind me to check over the whole estate once S90 [the fix] is live - call me cynical but I do not just accept a 3rd party's word that they have fixed something!"
So a serious bug affecting multiple branches for years.

Mr Green takes Mr Godeseth to a letter from Post Office's solicitors to the claimants' solicitors dated 28 July 2016. The letter asserts the Callendar Square bug "was discovered in 2005 and fixed in March 2006".  He then shows Mr Godeseth another Post Office solicitors letter which says: ""The Falkirk/Callendar Square issue was only known to have affected that one branch."

Mr Green invites Mr Godeseth to agree that these assertions are both wrong.

"On the face of it." replies Mr Godeseth.

The Receipts and Payments Mismatch bug

This was a Horizon Online bug, which Mr Godeseth says in his witness statement started in September 2010.

The problem is, Mr Godeseth had got his information about this bug from Gareth Jenkins, which as we have previously seen, may not be good information to put in a witness statement. Mr Green double-checked:

"So your evidence is that it arose in September 2010, yes?"
"Yes."
"And Fujitsu identified it in September 2010?"
"I have not checked the detail, but I'm comfortable with that, yes."

That was unwise. Even I know about this bug's history and I'm not even Horizon's Chief Architect (though I was paying attention during the first trial).

Mr Godeseth is taken to an internal Post Office document about the receipts and payments mismatch bug which was written between 29 September and 4 October 2010: "Discrepancies showing at the Horizon counter disappear when the branch follows certain... steps but will still show within the back end branch account ... impacting circa 40 branches since migration [from legacy Horizon to Horizon online] ... with an overall cash value of... £20K loss."

Furthermore: "At this time we have not communicated with branches affected and we do not believe they are exploiting this bug intentionally."

And: "We have asked Fujitsu why it has taken so long to react to and escalate an issue which began in May. They will provide feedback in due course."

Although he had been given a chance to challenge the information in his own witness statement a minute or two earlier, Mr Godeseth had chosen to say he was comfortable with it. He was asked about it again. He agreed that his witness statement suggesting the receipts and payments bug started in September 2010 was materially wrong.

Mr Godeseth's witness statement also suggested that 60 (later 62) branches had been affected by this bug. He wasn't very clear on how he had arrived at this figure, but it appeared to come about during discussions with Gareth Jenkins.

Mr Green drew Mr Godeseth's attention to a spreadsheet dated 16 July 2010, well before Mr Jenkins suggested to Mr Godeseth that the number of branches affected by this bug was in double figures:

"If we look please at the "Affected Branches" spreadsheet for a moment please and if you look at line 7, do you see there minus 777 in July 2010?"

Mr Godeseth did.

The local suspense account bug

According to Mr Godeseth's witness statement:
"As a result of some changes that were made on 3 July 2011 to the Horizon archiving strategy relating to Stock Units that had been deleted in a branch, any branch that deleted a Stock Unit at the end of 2010 which had a local suspense transaction in that Stock Unit before it was deleted were left in the table used for constructing the branch trading statement."
It transpired that this section of Mr Godeseth's witness statement had essentially been copied and pasted from a document written by Gareth Jenkins. According to Mr Godeseth:
"The bug was discovered in January 2013 when two Subpostmasters, who experienced the largest discrepancies, raised the issue with Post Office. I understand from Gareth Jenkins that Post Office could see the impact of the problem in their back end system and wrote off the discrepancies."
This resulted to the following exchange, led by Mr Green:

"The point there is that although it had persisted for years not months, it was SPMs who drew it to Post Office's attention and when they did, then Fujitsu saw it and was able to try and correct it?"
"We became aware of it in 2013, from my understanding."
"Yes, because it was raised by SPMs?"
"Correct."
"And Post Office had known about it since 2012?"
"I can't comment on that."

The Dalmellington bug

This time Mr Green stopped teasing and dealt with the Gareth issue up front:

"This was also known as the Branch Outreach issue and again here you are heavily reliant on Mr Jenkins who you refer to at paragraphs 55, 57, 58 and 61 of your witness statement, is that fair?"
"Correct." replied Mr Godeseth.

Having got that out of the way, we found out more about the Dalmellington bug from an internal Fujitsu document. It affected 65 branches and had two fixes, one issued in 2010 and another issued in 2011.

Neither did the job. The same internal document listed several reports of the bug resurfacing in several branches between 2011 and 2016. The document notes a 2016 fix to be released with Post Office's approval. The document also notes that between 2012 and 2014 the Post Office didn't raise the issue with Fujitsu.

"So the point that someone was making at Fujitsu" said Mr Green, "was that Post Office had not raised the matter with Fujitsu so that it could be corrected earlier.... isn't it?"

Mr Godeseth refused to comment.
"What do you think it looks like?" ventured Mr Green.
"It looks like that." conceded Mr Godeseth.

Mr Green went on to suggest that because the Post Office hadn't told Fujitsu about the reoccurrence of this bug, no root cause analysis was being done on it.

Mr Godeseth replied: "I think these incidents came from us going through the audit trail to identify where there had been duplicate pouches."
"So SPMs were suffering duplicate pouches over a number of years due to a known problem with the system and Fujitsu was not informed over that period of years until 2015?"

Mr Godeseth wasn't sure, so Mr Green tried a wider point:

"It does not show that any supposed countermeasures to ensure the robustness of Horizon as a system were working, does it?"

Mr Godeseth answered:
"Unfortunately this particular error is not subject to receipts and payments problems because it could be... it could be user doing something twice, it's the bug had the effect of making it look as though a user was simply doing something multiple times."
You have to admit it's getting better

Before Mr Godeseth left the witness box he was asked if he thought Horizon as a system had improved over time. Mr Godeseth asserted he believed all systems improved over time. He was asked if he thought that specifically about Horizon. "Yes." said Mr Godeseth.

Mr Green took him to a recent issue which arose when a requirement to remove mobile phone 'Top Up' options associated with Orange and T-mobile, now they had been bought by EE.

Unfortunately:
"there was a human mistake made when undertaking the change to the reference data for these products. The 'front-end' reference was correctly removed. The 'back-end' reference was incorrectly removed at the same time... This would have shown as a mismatch that the branch had to accept."
This was not just an example of a discrepancy caused outside of a branch directly affecting a branch's accounts, but, according to Mr Green it was a recent one which shows us how "how branch data is incorporated into transaction data and accounts for which a mismatch for which an SPM may be held liable can arise."

Mr Godeseth pointed out it would be picked up by Horizon's own monitoring systems.

Mr Green moved on to another problem caused by the introduction of Smart ID (known internally as EUM) across the branch network, designed to "allow a given EUM user to lock a counter and then log on to another counter."

The internal report about this change noted EUM "has caused incidents to occur in the live estate ... if the active user session rolls over the current stock unit, and a locked session on another counter attached to the same stock unit is then resumed."

Mr Green noted that as of 22 January 2019, the bug had "impacted 19 branches to date over a 6-month period."

Mr Godeseth agreed it was a new problem, but Mr Green had a newer one. This was dated Feb 2019 and was caused by: "a data cleanse on obsolete IDs". When it came to "Official Postage" which was due to end, there was a problem:
"When a branch was processing a claim for Official Postage this was increasing the cash on hand figure rather than decreasing which could potentially result in the branch having a discrepancy if not identified. NBSC received further calls and it became apparent that the accounting sense being applied at the counter was positive value not negative value. We spoke to a couple of contacts at Fujitsu but they couldn't explain why this was so"
Oh dear. Impact? "As of [2 February 2019] 720 branches are impacted including the live kit in NBSC ..."

Mr Godeseth didn't know about that one.

Where do we go from here

The brutal dissection of Mr Godeseth's knowledge and witness statement(s) was a bleak reminder of Richard Roll's trammelling at the hands of Mr de Garr Robinson. Cross-examination of knowledgeable people at the hands of a skilled QC is a vital part of the justice system and can have a dramatic effect on the understanding of a situation. You just wouldn't want to be on the end of it.

Nonetheless, over four days of questioning by Patrick Green for the claimants, we discovered there was a possibility the Post Office may have used incomplete evidence to prosecute people for theft and false accounting, a litany of Horizon errors revealed during Angela van den Bogerd's cross-examination, the bizarre approach the Post Office's debt collection department takes when it comes to branch discrepancies and the previous week, during the cross-examination of claimant witnesses, we found out about the sheer scale of the tinkering in branch accounts by Fujitsu engineers revealed by Richard Roll.

But Mr Godeseth's cross-examination was something else altogether. It contained an admission of a smoking gun - a branch discrepancy which the evidence suggests was more than likely caused by a Fujitsu engineer, being pinned by the Post Office on a hapless Subpostmaster.

There are two more Post Office witnesses to be cross-examined, but as things stand we are in limbo, waiting for the hearing of the Post Office's recusal application.

Stay tuned for more developments. Or come along to the recusal hearing on Wed 3 April at the High Court's Rolls Building. All welcome.

*****************************
Please feel free to forward the link to this blog post. The more people who read it, the more people find out about what is the biggest trial going through the UK courts right now.

Please also consider a donation to keep the postofficetrial.com website going by choosing an amount and clicking the button at the bottom of this post. Contributors who donate £20 or more will be added to the secret email list. Thanks!


Choose an amount

Friday 22 March 2019

Horizon trial: Je recuse

Bovvered?
There was bemusement all round as the implications of yesterday's recusal application by the Post Office reverberated around the court. After the judge had risen for the final time, I cheekily asked a group of highly paid professional people (and some journalists) to give me their instant reactions: "Desperate", "astonishing", "a delaying tactic", "bizarre" and, intriguingly, "last throw of the dice" came back.

Both legal teams knew what was coming - just - but the judge only found out at 1.55pm as he was about to walk into court. During the afternoon session he rose twice as everyone tried to work out what to do, which meant the journalists and court attendees standing around me had plenty of time to try to make sense of what they were hearing. No one could quite believe it.

This is the application:

"The Defendant [the Post Office] applies for an order that, due to the apparent bias of the Hon. Mr Justice Fraser, as manifested in the judgment handed down in this matter on 15 March 2019, the Hon. Mr Justice Fraser shall no longer be the Managing Judge of these proceedings. The Defendant further applies for an Order that the ongoing trial in these proceedings, over which the Hon. Mr Justice Fraser is currently presiding, should be adjourned."

As I wrote in "Going Postal" yesterday, the Post Office seems particularly incensed that the judge has some choice words about the credibility of the Post Office's witnesses and the Post Office itself. In a witness statement attached to the application, a Post Office solicitor believes the common issues trial judgment:

"contains a great deal of critical invective directed at Post Office, none of which is relevant to the determination of the common issues. That... creates a clear impression that the judge has not behaved impartially. The same can be said for those parts of the judgment which harshly criticised Post Office's witnesses on matters irrelevant to the common issues."

Patrick Green, QC for the claimants immediately stood up and said in response (and I paraphrase) "you'll need to be a bit more specific than that, chummy" - which takes us to the next step, the Post Office spelling out precisely where fair criticism becomes "harsh" criticism and "invective", and why that is evidence of bias.

What is recusal?

It's like a mix between excusal and removal. Oh you mean actual legal definition? Check out this helpful Law Society Gazette article, which the wonderful Kim dug out for me:

"The doctrine of judicial recusal dictates that a judge may recuse himself from proceedings if he decides that it is not appropriate for him to hear a case listed to be heard by him. A judge may recuse himself when a party applies to him to do so. A judge must step down in circumstances where there appears to be bias or ‘apparent bias’."

The article also notes the test of bias:

"if a fair-minded and informed observer, having considered the facts, would conclude that there was a real possibility that the judge was biased, the judge must recuse himself."

The oddest thing about the process is that the person who decides if the presiding judge is biased is the presiding judge. If he decides he's not biased then the Post Office can appeal, presumably to his boss, whoever that is.

Alongside the application the Post Office corporate machine went on a PR offensive, sending out a media release describing the recusal application as, well, a recusal application, and adding that they were acutely aware of its significance. Yuh-huh. They still aren't yet rejecting the judgment wholesale, saying:

"We will continue with our programme to improve the way we work with postmasters and are urgently considering how we can accelerate this work and address the criticisms about our operations raised in the judgment."

Alan Bates, founder of the Justice for Subpostmasters' Alliance was in no mood to mince his words, bellowing:

"It seems that Post Office’s current move is little more than what, in American Football, is known as a ‘Hail Mary’.  Having already thrown the ‘kitchen sink’ at the case, what is left for Post Office, nailing the doors of the court shut?"

What does everyone else think?

Most people following the case had a hot take. Chris Dubery, tweeted:
Darius Connelly, whose wife Deirdre is a claimant in the case tweeted:
Steven Murdoch, an Associate Professor and Royal Society Research Fellow tweeted:

One correspondent via email said: "Post office are quite happy to trust the judicial system when going after subpostmasters.  Paula Vennells once said something along the lines of 'after all it was the courts that convicted them.'  I've seen first hand the effects of their bullying behaviour and can't believe their latest stunt."

Another: "So its more of the same.  We were hounded out of our money and lively hoods by an organisation using bully boy tactics with no redress and now the judge gets the same treatment.  All this with public money and backed by the Government.  Why am I not surprised."

And: "If you don’t like the score, shoot the ref."

Finally Mark emailed to say "WHAT ABOUT KIM?", which was very funny, but unless you knew that I had Kim lined up to sub for me whilst I was away w/c 8 April, you may not have seen the joke. I have spoken to Kim, Mark. Don't worry,she is fine.

Onwards

The application to recuse does several things (even if it is unsuccessful), all of which favour the the Post Office.

First, it buys them time to work out what their strategy is going to be in the face of the first judgment, which given they don't have one yet, suggests it was a big surprise.

It continues to ramp up the litigation cost for the claimants.

Most importantly it means any eventual settlement in favour of the claimants will be less than it might have been before the recusal application was put in.

I tweeted something similar yesterday and a few people have suggested I am wrong about this, but I had a lengthy conversation with a couple of legal types near the Rolls Building yesterday, who helped me with my working out. I may have got the wrong end of the stick, but this is how I understand it:

the claimants are funded by a group who, if successful will want a return on their risky and very large investment.
The Rolls building and an inquisitive bollard

If the claimants are successful, the funders will get their investment back, plus a multiple, which could be anything between 10% and 30%.

Even if the claimants are succsessful and awarded ALL costs, the multiple has to come out of the damages. Damages which would otherwise be going into the claimant Subpostmasters pockets (or their creditors).

Now - my legal contacts told me - just because a trial drags on, or one party drags it out, there is no scope for punitive damages. The claimants have made a claim, and the amount they can claim (if successful) will be decided according to a formula which will be agreed once all the trials are concluded and the judge’s findings are handed down.

There is no scope in that formula for punitive damages or a reflection of the amount of time taken to complete the litigation.

So if the litigation was decided in the claimants favour now - they would receive roughly the same sum as they would in a year or two years' time.

But of course the costs keep growing, and so does the share of any settlement for the claimants which has to go to the funders.

Obviously the Post Office’s application to recuse the judge delays the trial (possibly for several months), and that ensures the claimants will get less money (net) if they win.

It also raises the possibility that if the Post Office successfully manages to drag the litigation on to the point at which the multiple on the funders’ costs would exceed the likely or possible damages, the claimants, if they win, could receive nothing, because their cash would be swallowed up by the funders’ multiple.

A former barrister got in touch after I tweeted all this yesterday and suggested I was being a touch "pessimistic". Well, I am just repeating what I was told, but as I say, I may have got the wrong end of the stick. If anyone is the final authority on this - please get in touch!

So is J Fraser about to get the heave-ho?

Nobody knows. Another lawyer I met (they're everywhere) said part of the effect of this recusal attempt might lead the judge to be exceptionally careful with his language in his judgments going forward.

I hope he continues to criticise where he sees fit.

Finally Mark Baker, described in the Common Issues trial judgment as "redoubtable" by the judge (which I presume was a compliment directly linked to his tenacity rather than any bias) tweeted:
The application hearing will be on 3 April 2019 in open court. The judge may well reserve his judgment so there might not be a result that day. But as one lawyer said to me - "make sure you're there!"

Oh yes. I will be.

***************************

Please feel free to forward the link to this blog post. The more people who read it, the more people find out about what is the biggest trial going through the UK courts right now.

Please also consider a donation to keep the postofficetrial.com website going by choosing an amount and clicking the button at the bottom of this post. Contributors who donate £20 or more will be added to this secret email list. Thanks!



Choose an amount

Thursday 21 March 2019

Horizon trial: day 8 - Going Postal

Post Office v Justice Fraser
What a week. A blistering judgment, revelation after revelation, and then this: the nuclear option.

We were expecting to witness the final session of Torstein Godeseth's cross-examination in court. Mr Godeseth is the chief architect for Horizon at Fujitsu. He'd spent the morning being taken through various serious Horizon errors, including a financial discrepancy mistakenly created by a Fujitsu engineer who was wandering around in a Subpostmaster's terminal without permission.

During a bungled attempt to fix a software glitch, the engineer miskeyed a digit, causing a $1000 discrepancy in a forex transaction. It wasn't picked up by Fujitsu at the time. It was only found during this litigation whilst the claimants' legal team were preparing for the Horizon trial. When it happened the Post Office held the Subpostmaster liable for the loss. More on that tomorrow.

Right now we're still reeling from the Post Office's attempt to have the judge recuse (sack) himself on grounds of apparent bias. It seemed to come as a surprise even to the Post Office's own QC, Anthony de Garr Robinson.

Mr Justice Fraser: in the firing line
At 2.02pm, after the usual lunch break, the judge walked into court and said: "I have received an application,
 Mr De Garr Robinson.  Do you know about it?"

Mr de Garr Robinson replied: "My Lord, I know that there has been an
 application, that is almost all I know."

The judge explained: "It is
 an application for me to recuse myself as being the managing judge in these proceedings which means
 effectively.... for this trial to stop.
 Although the application says "adjourn the trial",
 I think it really means start it again with another
 judge.
"

To say this is beyond extraordinary barely touches on it. The judge rose to give Mr de Garr Robinson a chance to try to work out what was going on. He returned ten minutes later to be told that Lord Grabiner, not Mr de Garr Robinson, would be acting for the Post Office in the application.

Lord Grabiner is One Essex Chambers' Big Dog. But unfortunately no one had been able to contact him, and so Mr de Garr Robinson and Mr Patrick Green QC for claimants would duke out the pre-Application procedural arguments, as the judge effectively decided the date on which his professional intergrity was going to be attacked.

Calculated to derail

Mr Green rose to his feet and drew his Lordship's attention to a witness statement supporting the application signed by Andy Parsons, a partner at Womble Bond Dickinson, the Post Office solicitors.

Lord Grabiner: The Big Dog
The witness statement says J Fraser's judgment of the common issues trial, which was handed down on 15 March:
 "contains a great deal of critical
 invective directed at Post Office, none of which is relevant to the determination of the common issues.
 That... creates a clear impression that the judge has
 not behaved impartially. The same can be said for those parts of the judgment which harshly criticised Post Office's witnesses on matters irrelevant to the common issues."


Mr Green made his point:

"I have not seen very many recusal applications, but I have not seen a recusal application as lacking in particularity as this... what is a fair appraisal of the credibility of a particular witness and what is invective?... This application, which... is likely if
 not calculated to derail these proceedings, has to be
 spelt out with proper particularity before we are in a position to respond to it."

This made everyone pause as the implications for the next battle in this deep war of attrition began to sink in. The judge had a little think:

MR JUSTICE FRASER:  All right, well, just before I ask you
 for a response on that, Mr De Garr Robinson, what I'm actually going to do is I'm going to go and read the whole witness statement, because I think I quite clearly have explained I haven't had a chance to read it yet.

MR DE GARR ROBINSON:  Nor have I.

MR JUSTICE FRASER:  Well, you might then also benefit from
 having had a chance to look at it.

MR DE GARR ROBINSON:  Indeed.


The judge rose again, and after a little bit more to-ing and fro-ing, the plan was agreed. Mr Parsons' witness statement would be fleshed out so everyone could see exactly what he was objecting to in the Common Issues trial judgment, and then (presuming the claimants opposed the application) the judge would hear it in court on 3 April and decide if he was biased (all welcome).

And if either party didn't like his decision, it could be appealed.

What was immediately and abundantly obvious was that the Horizon trial would have to be shut down until the matter can be resolved. And so it was.

The atmosphere in court after the judge rose was one of disbelief. Experienced, highly-paid people were literally shaking their heads with their mouths open.

This is going to get brutal.

Reaction

I am going to do more work on this on Friday. I'll post up some of the reaction from people who have contacted me over text, email, Whatsapp, email, the contact form on this website, twitter etc [UPDATE: I did exactly that. Read it here]

I'll also make sure that by the end of Monday that there is a proper fisking of Mr Godeseth's cross-examination, which was stunning, but which will now barely register in the wider world. [UPDATE: I did that, too! Read it here] Today's full court transcript is here.

***************************

Please feel free to forward the link to this blog post. The more people who read it, the more people find out about what is the biggest trial going through the UK courts right now.

Please also consider a donation to keep the postofficetrial.com website going by choosing an amount and clicking the button at the bottom of this post. Contributors who donate £20 or more will be added to this secret email list. Thanks!



Choose an amount

Horizon trial: Application to recuse the presiding judge


The application and supporting witness statement can be read on scribd in or embedded below. The reason for the application is the "apparent" (as opposed to actual) bias of the judge. The evidence for it is the first trial judgment and the reasons behind are listed in parts 24 and 25 of the witness statement below. What a day in court this has been. The application will be heard on 3 April 2019.

Horizon trial: day 8 transcript

Here is the transcript for day 8 of the Horizon trial, held at the High Court's Rolls Building on 21 March 2018. The trial was halted by an application for the judge to recuse himself on grounds of bias.

The application will be heard on Wednesday 3 April 2019.


                                        Thursday, 21 March 2019
   (10.30 am)
              MR TORSTEIN OLAV GODESETH (continued)
            Cross-examination by MR GREEN (continued)
   MR DE GARR ROBINSON:  My Lord, we are both standing up.
       I wonder if your Lordship wishes to discuss housekeeping
       matters or shall we leave it until after Mr Godeseth has
       finished?
   MR JUSTICE FRASER:  Absolutely.  We will do it at the end.
   MR GREEN:  I'm grateful.
           Mr Godeseth, just to clarify, yesterday -- and to
       correct a point I put to you yesterday -- I suggested to
       you that PEAK 0175821 at page {F/485} was
       Horizon Online.  It's clearly legacy.
   MR JUSTICE FRASER:  Was this the one that because of the
       dates could have been either?
   MR GREEN:  It could have been either.  I will come back to
       that but I just wanted Mr Godeseth not to be worried
       that --
   MR DE GARR ROBINSON:  My Lord, it was 2009, in fact, not
       2010.
   MR GREEN:  I'm going to try and take this in a condensed
       way, Mr Godeseth, if I can, and see what we can agree is
       clear.  Can we just -- what I'm proposing to do is take
       you to the position under the legacy system separately
       so there's no room for confusion by me or you and then
       deal with Horizon Online after that and just briefly, if
       we look first please at {F/323}.  Now, this is
       PEAK 0130275.  Do you see that at the top?
   A.  I do.
   Q.  And the target date there is June 2006, but you can see
       that from the progress narrative that it's initiated
       in December 2005.  Do you see that?
   A.  I do.
   Q.  And if you come down into the yellow box and you look
       towards the bottom you will see 21/12/05 at 17.22.
   A.  I do.
   Q.  Can you see the figure of £18,000 there just below that?
       It says:
           "Contacted: contacted pm - pm states rolling over tp
       with no items, 0 figures showing but allowed to rollover
       and on the next rollover into bp there was a £18,000
       discrepancy gain."
           Do you see that?
   A.  I see that.
   Q.  So this is a problem that's caused a gain to the SPM?
   A.  Agreed.
   Q.  On the face of the Horizon system?
   A.  On the face of it, yes.
   Q.  And if we go over the page please to page 2 {F/323/2},
       in the middle box you can see 22 December 2005 at
       16.05.15 and if you come down to "Stock unit BB" halfway
       down, do you see that?
   A.  Yes.
   Q.  "Stock unit BB was rolled over in an effectively empty
       state.  The PM then declared the correct amount of cash,
       and adjusted the stock levels up to the correct volumes.
       This has resulted in a gain of approximately £18,000.
           "We are unable to correct the system figures safely.
       We can however provide accurate figures for what should
       have been in the final balance for BB, to enable POL to
       make the correction perhaps by using a Transaction
       Correction."
           Now, that's one of the ways of dealing with
       a situation like this, isn't it?
   A.  It would be, yes.
   Q.  "POL need to make a decision on whether they are able to
       correct the problem in this way, however we do not see
       any other alternative.  Corrective action should be
       taken before 11 January when the branch is due to
       rollover into TP10."
           Do you see that?
   A.  I do.
   Q.  So at that stage, what's been identified is that it is
       really probably for POL to make the correction using
       a transaction correction, and it needs to be done before
       11 January because that's the next trading period
       rollover, yes?
   A.  Okay, understood.
   Q.  And if we go over to page 3 please {F/323/3}, 3 January,
       David Seddon, he says there:
           "If we get to the problem before the office is
       rolled we are able to change objects in the messagestore
       to reset the stockunit back to the CAP (TP) rollover
       trailer.  The PM can then rollover.  PM should get
       a large shortage which cancels out the large gain.
           "We don't want to be having to do this as making
       manual changes to the messagestore is open to error and
       each time we have to seek authorisation from [Post
       Office Limited] to make the changes."
           Now, pausing there, can we agree that it was
       possible to change the objects in the messagestore in
       the way described here?
   A.  I have -- yes, that would be possible.
   Q.  And that doing so would cause a large shortage which
       would cancel out the large gain, as expressly stated
       here?
   A.  Yes.
   Q.  And you would agree that making manual changes to the
       messagestore is open to error?
   A.  It certainly is.
   Q.  And were you aware that each time it was done,
       authorisation was sought from Post Office Limited to
       make the changes, or was that not something you were
       involved in?
   A.  At this stage I was working elsewhere, so -- 2006, early
       2006 I would have been involved interfacing with Horizon
       but certainly not at this operational level.
   Q.  So pausing there, just to get that very precise for
       the court.  Early 2006, which is the date of this
       particular PEAK, you were interfacing with Fujitsu with
       regard to Horizon in your job at Post Office?
   A.  I'm pretty sure I would have been because Impact was
       going live in 2005, so I can't be -- but certainly
       I wasn't involved in operational issues of this type.
   Q.  So you're not really the person to ask about those
       authorities being sought?
   A.  Certainly not the -- Post Office would have given
       authority through other parties.  Certainly I --
       I wouldn't have been involved in giving authority for
       this sort of thing.
   Q.  Thank you.  Then can we move forward then please to
       {F/432}.  Now, this is 10/12/2007 and it is PEAK number
       0152014 and you will see that there's an OCP listed
       there, OCP 17510 and an OCR listed there, 17532.
       There's also an additional OCR listed below.  Do you see
       that?
   A.  I do.
   Q.  Now, just so you have a fair context to this question,
       Mr Godeseth, it's accepted on behalf of the claimants
       that we do see references behind some of these documents
       to subpostmasters being made aware of changes being made
       of this sort.  So just so you're clear, it's not being
       put to you that they were never made aware, but I want
       you to look and see if you would agree with what this
       PEAK shows.
           Let's go over the page please {F/432/2}.  If we look
       at the top of the page, the yellow box, halfway down it,
       do you see:
           "Am currently retrieving the messagestore for this
       branch [and] will then be inserting a new message on the
       counter to remove the effects of this."
           Do you see that sentence?
   A.  Yes, and I read this overnight.
   MR JUSTICE FRASER:  Sorry, can you tell me where it is?
   MR GREEN:  I'm so sorry, in the yellow box at the top,
       my Lord.
   MR JUSTICE FRASER:  Oh, the yellow box at the top.  I was
       looking halfway down the page.
   MR GREEN:  Sorry, my mistake.
           So this is one you have had the opportunity to look
       at overnight and it clearly states there that Andy Keil,
       who we see on the previous page, just to give you
       context {F/432/1}, 10 December 2007, 12.36, Andy Keil --
       or is it Andy Keil?
   A.  I don't remember him.
   Q.  Okay.  So he is saying -- go over back again {F/432/2},
       he is talking about a line written -- if we look at the
       top of that yellow box -- he is talking about a problem
       which is:
           "... due to a single SC line written for $1,000
       (£484) with no settlement in the middle of two RISP
       transactions."
           Can you tell the court what RISP transactions are?
   A.  I'm not -- RI is remit in; I can't remember what the SP
       bit stands for.
   MR JUSTICE FRASER:  And SC line?
   A.  Sorry?  SC would be "serve customer", my Lord, so that's
       a sale.
   MR JUSTICE FRASER:  So it looks as if it they are selling
       foreign currency?
   A.  This one is all related to selling foreign currency and
       there's a bit of confusion as to whether it is American
       or Canadian dollars.
   MR JUSTICE FRASER:  In the happy days when £484 sterling
       bought 1,000 of either US or Canadian dollars.
   A.  And there was a change in exchange rates the last --
       over the couple of days, yes.
   MR GREEN:  When we see halfway down that top yellow box:
           "Am currently retrieving the messagestore for this
       branch, we will then be inserting a new message on the
       counter to remove the effects of this."
           What's being considered there is making a change on
       the counter messagestore, isn't it?
   A.  Yes, my Lord.
   Q.  And halfway down you can see a double star, double
       asterisk followed by:
           "Again, this may also have caused a receipts and
       payments error ..."
           So they are trying to work out whether there's
       a gain or a loss at the counter and the amount, yes?
   A.  Yes, my Lord.
   Q.  And then if we go down to the bottom of this page,
       12 December 2007, Andy Keil says:
           "Worth noting that the branch did not have any
       issues with the mismatched transactions because this was
       fixed before they did the roll.  The branch is not aware
       of this and it's best that the branch is not advised."
           Now, at that stage you would fairly accept, wouldn't
       you, that although there are documents elsewhere that
       show that sometimes SPMs were advised, this shows an
       occasion when the SPM at this stage has not been
       advised, doesn't it?
   A.  I don't know that it shows that, but it's a fair
       inference from it.  I think I would say that Post Office
       were well aware of this and I would argue that it's
       a Post Office decision whether or not to tell
       a subpostmaster.
   Q.  So you are saying -- just clarify what you just said.
   A.  This was written by Andy Keil who is a Fujitsu employee
       and I think he is recording things as he saw it, but in
       the background to this there was a dialogue with
       Post Office, from my reading of the PEAK.
   Q.  Yes, so Post Office would have been aware of what was
       being done by Fujitsu in this respect?
   A.  Correct, my Lord.
   Q.  And then what seems to happen, if we go over the page to
       {F/432/3}, in the big yellow box at the top:
           "The counter problem which caused the first issue
       has been corrected by inserting a message into the
       messagestore, for equal but opposite values/quantities,
       as agreed with POL ..."
           Yes?
   A.  Yes.
   Q.  And that refers there to an OCP, 17510.  Do you see
       that?
   A.  I do.
   Q.  And if we could just look at that for a moment please,
       it's at {F/432.2}, "Write corrective bureau message for
       FAD 183227."
   MR JUSTICE FRASER:  Where are you reading?
   MR GREEN:  That's the heading of --
   MR JUSTICE FRASER:  The bold heading.
   MR GREEN:  The bold heading, my Lord, yes.  Then it says:
           "A single SC message ... was written in error on
       26th November ... selling 1,000 US dollars, with no
       corresponding settlement line.  To remove the effects of
       this message at both the branch and on POLFS, we will
       inset a new message to negate the effects of the
       original message."
           Yes?
   A.  Yes, my Lord.
   Q.  "Justification: If the change is not made in the counter
       messagestore (before the stock unit is balanced on
       Wednesday), the branch will have an unexpected gain of
       £484 (or thereabouts - [depending] on exchange rate),
       and a receipts and payments mismatch.  This gain would
       have to be resolved at the branch.  There would also be
       an inconsistency between the branch and POLFS to be
       resolved.  By correcting the problem locally, the branch
       may not be aware of the problem, and there will be no
       inconsistency between the branch and POLFS."
           Yes?
   A.  Agreed, yes.
   Q.  So this is the reason for doing it, and it is planned
       for 10 December 2007 at 5 o'clock with a duration of
       ten minutes.
   A.  Yes.
   Q.  And then the extra detail provided says:
           "The original message had Product [ID]: 5129,
       [quantity 1], Sale value~... 484, [quantity]~...
       1,000~..."
           The new message will have quantity --
   MR JUSTICE FRASER:  Quantity minus 1.
   MR GREEN:  Sorry:
           "[Quantity]:-1, SaleValue:-484, PQty:-1,000 with,
       other attributes (including exchange rate) as before."
           So the OCP is the front-end fix, isn't it?
   A.  Yes, this is going into messagestore.
   Q.  This is going into the counter messagestore at the
       branch?
   A.  Yes.
   Q.  And we can see that if we go over the page {F/432.2/2},
       because at "Change at location" we see "Counter"?
   A.  Yes.
   Q.  And when the location at which the change is effected
       for the back-end in an OCR, we tend to see location
       BRA01 or one of the terminals within Bracknell or --
       yes?
   A.  Understood.
   Q.  When we look at the relevant OCR, which is OCR17532
       which is on page {F/434.1}, we can see there that the
       change at location code is BRA01, as agreed already.
   MR JUSTICE FRASER:  Where are you looking?
   MR GREEN:  In the middle of the page under "Other details".
   MR JUSTICE FRASER:  Yes.
   MR GREEN:  And if we look under "Comments" we can see
       Andy Keil wrote at 12/12, he is updating the POLFS feed
       for the branch with a sale value of 1,014.73 and PQty of
       2,080.  Now, there's a difference between the updated
       value that we saw on 432.2, which was a $1,000
       transaction being inserted at the counter, and what we
       see at {F/434.1} which is just over $2,000 being
       inserted in the Post Office system, yes?
   A.  Yes.
   Q.  When we go back to the underlying PEAK at {F/432} --
       hopefully you've got hard copies as well?
   A.  I believe I have.
   Q.  And if you want to look at those, please take your time.
       If we look at page {F/432/3}, so the third page of that
       PEAK, on 14 December 2007 Anne Chambers, in the large
       yellow box at the top, towards the top of the page, she
       writes, in the second paragraph there:
           "The counter problem which caused the first issue
       has been corrected by inserting a message into the
       messagestore, for equal but opposite values/quantities,
       as agreed with POL."
           Which we have already looked at and there's the
       reference to the OCP:
           "As a result of this corrective action, the net
       effect on POLFS is zero, and POLFS figures are in line
       with the branch.  POLMIS received both the original
       message and the corrective message.
           "Once the problem was corrected, there should have
       been no impact on the branch.  However, it has been
       noted that the stock unit BDC had a loss of $1,000,
       which was generated after the correction was made."
           Now, that's of roughly the same amount as the
       difference between what was entered in the front-end and
       the back-end, wasn't it?
   A.  Yes.
   Q.  "We have already notified Gary Blackburn at POL ... this
       appears to be a genuine loss at the branch, not
       a consequence of the problem or correction."
           Would you accept that it's a fair possibility that
       the additional loss was in fact a result of the
       difference in the two corrections made by the OCP and
       OCR?
   A.  Having read this PEAK in more detail overnight then yes,
       clearly that is what appears to have been the case.
   Q.  Can we move forward now to the 2009 PEAK at {F/485}.
       This is PEAK 0175821 and this has a target date of
       22 February 2009, and you can see that if we come down
       to 19 February you can see there's an OCR reference
       there, OCR 21847.  And if we go down, just to give you
       context -- hopefully you have had a chance to look at
       this.  At the bottom of the page, 19 February 2009 -- we
       touched on this -- at 17.39.40, Catherine Obeng:
           "There are two sides to the problem relating to
       these transactions.  The first is where all five SC
       [transactions] missing core data as described in the
       above-mentioned KEL.  Second is absence of equal but
       opposite (ie settlement) lines.  See [another PEAK]
       PC0152014 for a similar problem and how problem was
       resolved.
           "For the first problem, I have used the TRT ..."
           That's the transaction repair tool?
   A.  Correct.
   Q.  "... to insert the missing data ie Region, Margin,
       Margin Product and EffectiveEXRate."
           Yes?
   A.  Yes.
   Q.  Now, when we look over the page, please, at page 2
       {F/485/2}, it is reasonably complicated to follow but
       just to trying to take out a few simple points from it.
       If we look at the second light blue box,
       20 February 2009, Garrett Simpson, this is -- he says:
       Following the investigation shown in another related
       PEAK he has analysed the:
           "~... TPS_POL_FS_Summaries_Incomp where article ~...
       ='USD'."
           This is another one where there seems to be an
       exchange rate issue, is that fair?
   A.  Certainly dealing with bureau is complicated by the fact
       that there are exchange rates, there are margins, there
       are quite a few components to the transaction.
   Q.  Indeed.  And we can see the difference between the two
       figures that have been recorded is 378.42 but the total
       non-zero value for the branch is £989.96.  So he says:
           "So I have not found the proper exception".
           What does "I have not found the proper exception"
       mean?
   A.  I think this may well be the one where there was -- it
       appeared to me to be confusion as to whether people were
       talking about American dollars or Canadian dollars.
   Q.  Yes.
   A.  And I think further on in the PEAK, he identifies that
       he has found the correct element to look at.
   Q.  Yes.  If we go -- well, can I ask you this shortly: it
       seems that there are two attempts to remedy this, is
       that fair?
   A.  I think that was the previous one.  Sorry, I think there
       were two components to fixing this, in that there was
       a change using the TIP repair tool to get the feed into
       POLFS correct and then there was a change to the
       messagestore to get the branch aligned.
   Q.  Yes.  There are two aspects to it, the first using the
       TIP repair tool and the second is the change to the
       branch messagestore?
   A.  And I would see that as a way of keeping the branch and
       POLFS aligned.  The repair had to go two ways.
   MR JUSTICE FRASER:  Can you just remind me which
       three-letter abbreviation are we using for TIP repair
       tool?
   A.  That is probably TRT -- no, sorry --
   MR GREEN:  It is TRT.
   A.  TRT on OCR17532, it says "Change type TRT" which I take
       to mean TIP repair tool.
   MR GREEN:  And your Lordship will also find a reference to
       it at {F/485} on the first page.
   MR JUSTICE FRASER:  I saw the reference, it is just we had
       been using TRT, although I know you mentioned yesterday
       that's what you called the TIP repair tool, I think.
   A.  I have forgotten, my Lord.
   MR JUSTICE FRASER:  All right, don't worry.
   MR GREEN:  And just for completeness, if we look very kindly
       at {F/485.2}, we can see that this is the OCP, 21918,
       this is the change to branch messagestore OCP and if we
       go over the page --
   MR JUSTICE FRASER:  Can we just go back a second.
   MR GREEN:  Sorry.
   MR JUSTICE FRASER:  Under "Extra detail", Mr Godeseth, is
       that just an explanation in prose of what you said were
       the two components to remedying it?
   A.  I'm reading that as saying that we tested what we were
       about to do on the live system.  We tested in a test
       system what we were about to do on a live system to
       ensure that this was going to work.
   MR JUSTICE FRASER:  Yes, but in other words, that includes
       the two components, does it?  In other words, are your
       two components together, as far as you can tell --
   A.  I can't tell from this I'm afraid, my Lord.
   MR JUSTICE FRASER:  You can't, okay, that's fine.
   MR GREEN:  My Lord, I think I can help Mr Godeseth with
       that, if that assists.
           If we look -- the OCP is normally for the front-end,
       the branch messagestore, and the OCR is normally the
       back-end and if we look at {F/485.1} we have 21847.  We
       can see the deletion of transactions and updating of
       transactions, extracts from some of the scripts used
       there, can't we?
   A.  Yes.
   Q.  So the OCR relates to the POLFS system records?
   A.  Correct.
   Q.  And that's 485.1, and then the {F/485.2} is the OCP for
       the branch messagestore?
   A.  Yes.
   Q.  And we can see there -- it's actually at the second half
       of this page, we can see Gaby Reynolds of Post Office
       Limited wrote --
   A.  Gaby Reynolds was Fujitsu.
   Q.  POL -- okay.  So it says "POL" is that she is on the
       Post Office account or ...?
   A.  Certainly the Gaby Reynolds I knew worked for Fujitsu
       and never worked for Post Office.
   Q.  Okay, very good.  So it seems to be between
       Gaby Reynolds of Fujitsu -- sorry, received by
       Gaby Reynolds of Fujitsu, from julie.edgeley@postoffice,
       on behalf of duty_manager@royalmail.com,
       27 February 2009 to Anne Chambers, who we see a lot in
       in these.
   A.  Yes.
   Q.  And it says:
           "~... malformed currency transactions ~...
           "As discussed, POL are happy for you to make the
       necessary system adjustments.  From speaking to Wendy,
       the manager in the branch, first thing on Tuesday
       morning (between 9am and 10am) is the quietest time for
       them.
           "I have advised Wendy that you will call her as you
       are about to start and as you finish.
           "Thank you."
           We can see over the page that that's then forwarded
       on {F/485.2/2}.  Or you've got a separate one from
       Nicola Watson, Post Office, saying that:
           "POL approve this change."
   A.  Yes.
   Q.  So effectively the agreement has come from two people at
       Post Office and we have also seen from this particular
       one that the SPM must have been made aware of it?
   A.  Yes.
   Q.  Because they're doing it at a time convenient for the
       SPM.  So that subpostmaster would have known that
       changes could be effected at the branch messagestore, or
       maybe not in those words, but certainly at the counter?
   A.  Certainly changes made to the messagestore.  I don't
       know whether we are discussing whether it was injected
       at the counter or in the correspondence server here,
       I can't remember the detail of this.
   Q.  Okay.  Let's move forward, if we may --
   MR JUSTICE FRASER:  Yes, before we do I would like to go
       back to {F/485.2/1} because Mr Green said he was going
       to help you with something and I'm not necessarily sure
       he did.
   MR GREEN:  My Lord, sorry.
   MR JUSTICE FRASER:  No, that's all right.
           As I understand it, OCP deals with the front-end, is
       that right?
   A.  It appears so, certainly, my Lord.
   MR JUSTICE FRASER:  As far as you know?
   A.  Yes.
   MR JUSTICE FRASER:  And OCR deals at the back-end?
   A.  That seems consistent, my Lord.
   MR JUSTICE FRASER:  Yes.  So when you were saying there are
       two components to this correction, did you mean one
       component was the OCP and one component was the OCR?
   A.  Yes, my Lord.
   MR JUSTICE FRASER:  But that will be the case in all
       corrections, won't it?  There will always have to be
       a correction at the front-end and the back-end.
   A.  No, not necessarily.  The TIP repair tool was used quite
       a lot to repair stuff which had gone through from the
       branches -- or from the counters fine but it was missing
       some attribute which POLFS or something else further up
       the line needed and so therefore it would fail
       validation at the point when it was hitting TIP and it
       would be parked until it was repaired.
   MR JUSTICE FRASER:  I see.  So your two components were one
       of them was the OCP and one of them was the OCR?
   A.  Correct.
   MR JUSTICE FRASER:  All right.
   A.  For this particular problem we needed to change in two
       places -- or they chose to change in two places.
       Possibly it could have been done in a different way.
   MR JUSTICE FRASER:  Understood.  And when you look under
       "Extra detail", the last sentence -- and I understand
       this relates to foreign currency which is why it is
       said:
           "The gain may not be precisely the same as the
       original loss ..."
           Would there have been another component as well to
       correct that, so far as you know, or do you not know?
   A.  It is tied up in exchange rates and margins, so there is
       a mechanism in place to account for changes in currency
       rates which would have to be replicated through this, so
       that it would be taken into account.
   MR JUSTICE FRASER:  And would that be another OCP then?
   A.  I don't imagine so.  I think it would all be dealt with
       in one go.
   MR JUSTICE FRASER:  In the same OCP?
   A.  Correct.
   MR JUSTICE FRASER:  Mr Green, over to you.
   MR GREEN:  Can we move forward now to 2010 and we've got
       PC0195561 which is 7 March 2010.
           Now, we have seen this before?
   A.  Sorry, I'm not seeing it yet, my Lord.
   Q.  Sorry, {F/590}.  Have you not seen this one before?
   A.  Sorry, it was --
   MR JUSTICE FRASER:  No, it wasn't on the screen, Mr Green.
   MR GREEN:  I'm so sorry.
   MR DE GARR ROBINSON:  It wasn't given last night either.
   MR GREEN:  This is the one that Mr Godeseth himself refers
       to at paragraph 58.8 as the balancing transaction that
       he is aware of.
           If we look in the yellow box below the second line
       of equals signs:
           "~... pm was trying to transfer out 4,000 pds.  The
       system crashed.  Pm was issued with 2 x 4,000pds
       receipts."
           Yes?
   A.  Yes.
   Q.  So this is one you are familiar with?
   A.  Yes.
   Q.  And if we go to page 3 {F/590/3}, 10 March 2010 at
       8.51.33, almost the top yellow box:
           "After discussion with Gareth Jenkins, the suggested
       correction is to negate the duplicate transfer out by
       writing 2 lines to the BRDB_RX_REP_SESSION and
       BRDB_RX_EPOSS_ TRANSACTIONS tables ..."
           Yes?
   A.  Sorry, I missed -- sorry, could I just confirm that this
       does relate to the one use of BT?
   Q.  This is -- if we have a look at your witness
       statement --
   MR JUSTICE FRASER:  Do you have your hard copy witness
       statement?
   A.  I do.  Which?
   MR JUSTICE FRASER:  I think it might be the one you are
       being asked about at {E2/1/16}, but if that's an
       incorrect reference, Mr Green will tell you.
   MR GREEN:  If we look at paragraph 58.8 {E2/1/17}, you say:
           "The PEAK incident ticket raised in relation to the
       BT ..."
   A.  Is 1955 ~... yes.
   Q.  Yes?
   A.  Yes, my Lord.
   Q.  So who have you got this information from about this?
   A.  I will have got this information -- this information
       will have come from the guys who knew about the PEAK,
       knew about the balancing transaction, the occurrence of
       it.
   Q.  So who were the guys?
   A.  I can't remember, but I got a lot of information about
       this it from Steve Parker.
   Q.  From Steve Parker?
   A.  So I imagine it came from -- via his team.
   Q.  So you got the information from Steve Parker and you
       imagine it came from his team?
   A.  Yes.
   Q.  And let's go back to the PEAK, if we can, {F/590/1}.  We
       have seen that it is basically doubling the £4,000 of
       receipts and we have moved on to page 3 {F/590/3} and
       we've got the discussion with Gareth Jenkins at the top
       suggesting writing two lines to the transaction tables,
       yes?
   A.  Yes, my Lord.
   Q.  And those tables are tables which record transaction
       data, aren't they?
   A.  Yes.
   Q.  And there it says:
           "An OCP approved by POL will be needed."
           It says:
           "[It] should be done using the Transaction
       Correction Tool.  An OCP approved by POL will be
       needed."
           Yes?
   A.  Yes.
   Q.  Do you know of any reason why the £4,000 difference
       couldn't have been corrected by a transaction correction
       being issued?
   A.  Yes, because this was a double entry -- this is on the
       new system.  This is not messagestore, so the only way
       to actually change the branch database so that I had two
       entries was to insert a record using the branch
       transaction tool, or the balancing transaction tool.
       Sorry, that is the only -- that is the chosen way to do
       it, it was a way that we had anticipated dealing with
       this type of problem.
   Q.  That's the way that had been planned for?
   A.  We obviously planned never to do it, but this was the
       contingency should a problem arise in the branch
       database which left us with an unbalanced transaction,
       ie one side had been written but the other side had not.
   Q.  If we can go please for {F/590/9}, so page 9 of the same
       PEAK, 22 April 2010 at 15.35.[54].  You are reasonably
       familiar with this PEAK, I think?
   A.  Yes.
   Q.  What we see in this box if I can just summarise it, is
       there had been two transactions with the same JSN number
       and the second transaction should have been rejected as
       a duplicate, but in fact, the data from the second entry
       had in fact been committed to the database when it
       shouldn't have been, and that was a bit puzzling.  Is
       that a fair summary?
   A.  I can agree it's all puzzling.  I'm not sure that this
       would be a totally accurate reflection of what happened.
       This is an interpretation put together by a developer
       who is investigating the problems and the --
   Q.  Well, if -- sorry.
   A.  The log files will still exist.
   MR JUSTICE FRASER:  Well, let's go on what's on the face of
       the page.  As I understand it, your answer is that
       you're not sure that what's on the screen is totally
       accurate, is that right?
   A.  Correct, my Lord.
   MR JUSTICE FRASER:  Let's now work on the hypothesis that it
       is, and Mr Green, continue.
   MR GREEN:  On that basis, what this shows is that there were
       two entries with the same JSN number presented.  The
       second one should have been rejected because the
       database should not accept data from duplicate JSN
       numbers --
   A.  Yes, my Lord.
   Q.  -- but in fact what's being said here is the data had
       nonetheless been committed to the database when it
       should not have been?
   A.  That's what I cannot understand, because the Oracle
       database -- since the JSN is part of the primary key
       into the message table, I cannot see how an Oracle
       database would allow that to happen.
   MR JUSTICE FRASER:  Because the fact it has accepted one
       means it wouldn't accept the other?
   A.  Correct, that's the whole --
   MR JUSTICE FRASER:  That's how it is supposed to work.
   A.  The primary key in the Oracle database is there to say
       "this is the one and only record" and trying to
       overwrite that would be rejected by the branch database.
   MR JUSTICE FRASER:  But if this is right -- and this is
       obviously something I will decide in due course after
       hearing the experts, but if this is right, that doesn't
       appear to have happened.
   A.  That would have to be a bug in Oracle, and it was
       certainly not the bug that I was looking -- it was not
       the bug that I was looking at for the red alert.
   MR JUSTICE FRASER:  Yes, all right.
   MR GREEN:  This is a different one?
   A.  If this in fact happened as written down here then it
       would have had to have been a bug in Oracle and
       I certainly don't remember any such bug.
   Q.  Let's move forward if I may --
   A.  Sorry, could I just bring your attention to the line
       which says:
           "I have requested for the journal table dump, to
       check whether duplicate JSN entries exists in the table.
       But from the DB dump, I couldn't find any duplicates."
           I see that as highly significant.
   MR JUSTICE FRASER:  Can you just tell me where that is?
   A.  Sorry, it is just above 2010-03-05 -- sorry, 15.05.
       15.05.18, it is the two lines just above that.
   MR JUSTICE FRASER:  I'm struggling, I'm afraid.
   MR DE GARR ROBINSON:  My Lord, it is halfway down the big
       yellow box.
   MR JUSTICE FRASER:  I was going to say, let's start with
       a colour.
   A.  Sorry, I'm just seeing all of this in green.
   MR JUSTICE FRASER:  Okay.
   A.  So if we go down to the line --
   MR JUSTICE FRASER:  Which box?  The big box at the bottom?
   A.  It is the big box which starts on -- 15.35.54.
   MR JUSTICE FRASER:  Got that.
   A.  If we carry on down, the next line with a date on it is
       2010/03/02, two chunks up from that it says:
           "I have requested for the journal table dump, to
       check whether duplicate JSN entries exists in the table.
       But from the DB dump, I couldn't find any duplicates."
   MR JUSTICE FRASER:  And then "Even I have requested ..." is
       that "OSR logs"?
   A.  Yes, my Lord.
   MR JUSTICE FRASER:  And OSR stands for?
   A.  It is online service routing, it is part of the BAL.  It
       is effectively a synonym for the BAL.
   MR JUSTICE FRASER:  And then does it say:
           "I didn't get any clue on the request ... I could
       only able to see the following in the osr.log file."
   A.  Yes.
   MR JUSTICE FRASER:  And do I take it that the code -- or the
       language that follows after that is what the OSR log
       file would say?
   A.  The OSR log is saying that there was a basket settlement
       transaction, it has timed out but is executing, so there
       was an attempt to write a message to the BRDB, but it
       has timed out.
   MR JUSTICE FRASER:  And am I right that everything on those
       two and a half lines that starts with 2010-03-02 appears
       to be an extract from the OSR log?
   A.  That -- yes.
   MR JUSTICE FRASER:  From the use of the words "The
       following", in the preceding --
   A.  Yes, that's a fair summary, my Lord.
   MR JUSTICE FRASER:  Thank you very much.  Mr Green.
   MR GREEN:  Could we just go over the page to follow that
       through to give you context {F/590/10}.  You can see
       that what's being looked at is -- you see the words
       "I suspect" at the top:
           "I suspect there must be something gone wrong with
       this request ..."
           Identifying the request:
           "~... unfortunately no clue on this.  I am not sure
       why this might have happened.  Normally, since this is
       a retried one it should have failed at the Journal
       filter stage."
           So there's a retried transaction which should have
       failed at the journal filter stage.
   A.  I think we're getting into fairly complicated territory
       here, my Lord.
   MR JUSTICE FRASER:  I think the whole case is fairly
       complicated territory.
   A.  Sorry.
   MR JUSTICE FRASER:  But I think I can follow it.
   A.  Fine.  It is quite a normal situation -- or it will
       happen that you will get the same message come from the
       counter to the BAL, because something has gone wrong if
       the counter is retrying whether it's a manual
       button-press to say please retry this, or it is an
       automated press, the same message will come up to the
       BAL, the BAL will say: I have had this message before,
       the answer last time was this, it will send the same
       response back down and that should take care of
       situations where you are having to retry because
       something has simply failed on the comms channel.
           In this situation the symptoms, as I'm reading them,
       are that because there was a bug in Horizon and this is
       in the new system, it was pretty early days of the
       Horizon Online, because there was a problem, we had
       something coming through which got through the journal
       filter but then failed at the branch database and so
       therefore, as far as the branch database is concerned,
       it has not happened.
   MR JUSTICE FRASER:  I understand that.  I think what
       Mr Green is putting to you is that this shouldn't have
       got past the journal filter.
   A.  It's a bug, so certainly the way that the system should
       have worked, a JSN -- the same JSN coming up would be
       just a simple repeat of the message for -- because there
       was some sort of glitch.
   MR JUSTICE FRASER:  Yes.  Over to you, Mr Green.
   MR GREEN:  And just to clarify a small point on that: is it
       fair that this box read as a whole shows that the person
       looking into it was surprised that the second lot of
       data had got into the BRDB, but there appeared to be no
       duplicate JSN entry found in the JSN dump?
   A.  My Lord, you couldn't have two records in the Oracle
       database with two JSNs unless there is an Oracle bug.
   Q.  So that's what they were looking for because the
       duplicate -- there were two lots of data had been
       entered in, so they looked to see if there had been
       a duplicate JSN entry to try and find that out?
   A.  Somebody may have been trying to look for a duplicate
       JSN entry.  I can't really comment on what the guy was
       doing at the time who was trying to investigate.  I am
       simply asserting that barring an Oracle bug, which would
       have been huge, you cannot have two entries with the
       same JSN.
   Q.  Let's move forward please now.  You will be pleased to
       know I think we only have two more PEAKs to go just by
       way of illustration.  {F/611} please.
           So this is just moving forward in time, we are now
       in April 2010, so this is a month after the acknowledged
       balancing transaction that you have referred to in your
       statement, yes?
   A.  Yes, my Lord.
   Q.  And we can see the problem there in the summary is FAD
       code "314642 unable to rollover".
   A.  Yes, my Lord.
   Q.  And this is PEAK 0197592 and if we look just below the
       second equals sign in the box you will see:
           "When rolling over and doing branch trading
       statements site gets message - unable to connect to data
       centre."
           Do you see that?
   A.  I do.
   Q.  And then there is a discussion that we see at 11.42 at
       the foot of {F/611/2}.  This is Anne Chambers and if we
       go on to the next page {F/611/3}, you can see the third
       paragraph she says:
           "The bal osr log shows an exception while executing
       statement insertOpeningBalanceForRollover ...
           "I suspect this may be because there is already
       a single entry in BRDB_SU [stock unit] _OPENING_BALANCE
       for DEF TP 12 BP 1 [trading period 12, balancing period
       1] inserted during migration.  The entry is for cash,
       zero value.
           "I'm wondering if this branch could be sorted out by
       changing the TP in BRDB_BRANCH_INFO (this would have to
       be done by OCP by development/ISD).  The PM already has
       several printed copies of the TP1 11 BTS. From the logs,
       I can see that suspense and cash/currencies awaiting
       collection are all zero -- so no office opening figures
       are required for TP 12??
           "This needs looking at urgently for a workaround for
       this branch, and longer term to see if this can be
       avoided at other migrating branches."
           So that's the context, isn't it?
   A.  Yes.
   Q.  If we look at 14 April 2010 at 1 o'clock and 9 seconds,
       you see Gareth Jenkins again:
           "I've had a look at this PEAK and agree that we need
       an OCP to tidy up BRDB to ?un-stick? this Branch.  Note
       that what I am proposing here is slightly different from
       what Anne has suggested above.
           "What we need to do is the following:
           (I know the SQL is wrong, but BRDB Host team can
       correct it and fill in the gaps)."
           So just pausing there, he doesn't seem to be going
       to use the template SQL code that had already been
       prepared at this stage, because he is going off piste
       a bit in some sense here, isn't he?
   A.  He is -- the balancing transaction tool is designed, or
       is intended to introduce the balancing part of a double
       entry item.  This is not, in my mind, a double entry
       item.  This is the opening balance for a trading period,
       or for a period --
   Q.  Yes.
   A.  -- and there is no second entry to it, there is nothing
       to balance that against.  The only thing you could
       balance it against would be the closing balance of the
       previous thing.  The way the system is designed, this is
       a number in a table which represents the opening balance
       of a particular period.
   Q.  Yes.  So he is using SQL code to set that correctly so
       that the problem we have seen above can be overcome at
       rollover?
   A.  Technically what is about to happen is an item is going
       to be deleted under careful -- so an entry in the branch
       database is about to be deleted.
   Q.  Yes, the --
   A.  It is not transaction data.
   Q.  Yes, I think there are two things that are happening,
       just to be absolutely precise.  One is an update to
       a stock unit entry, yes, which we see at the fourth
       paragraph at the top of the page?  There's there are
       a single entry inserted during migration, this is for
       cash zero value, so it looks like there's a stock unit
       entry being updated in number 1 and then a deletion to
       that opening figure of cash, is that right?
   A.  That's what that's saying, yes.
   MR JUSTICE FRASER:  And the second one is in respect of
       trading period 12, I think.
   A.  The second one is to remove an item which is saying it's
       the cash position in trading period 12, yes.
   MR JUSTICE FRASER:  The first one is in respect of trading
       period -- is setting it to trading period 11?
   A.  Yes, I must admit I hadn't sort of picked up on that
       nuance, but that would be saying that this is
       effectively a flag to say, or an indication as to
       which -- where you're at with that particular stock
       unit.
   MR JUSTICE FRASER:  I'm not making any findings or anything,
       but it's not exactly a nuance, is it, if they're two
       different trading periods?
   A.  Oh, sorry, no.  I must admit I had missed in my sort of
       reading through this that there was a problem with
       a stock unit being flagged as being in period 12.
   MR GREEN:  Can we move on now please to {F/594}.  This
       is October 2010.  It is PEAK number 0195962.
   MR JUSTICE FRASER:  I wonder if we could keep the murmuring
       backstage down a little bit please.  Carry on, Mr Green.
   MR GREEN:  Thank you very much, my Lord.
           This is later on in 2010, and if we look halfway
       down underneath Gareth Seemungal's name we can see that
       it appears to be created -- or the Impact statement is
       set at August 2010 and we can see underneath from the
       progress narrative that the actual PEAK is opened
       in March 2010.  Do you see those two dates?
   A.  I do.
   Q.  So this spans across the period we have been looking at,
       yes?  And what's said there is:
           "The proposed fix would correct and update the BRDB
       transaction correction tool templates, making it less
       likely that mistakes will occur when SSC are trying to
       resolve problems with transactions in BRDB."
           Yes?
   A.  Yes.
   Q.  And if we go down to Cheryl Card, 12 March 2.04.46:
           "The Transaction Correction tool has now been used
       in live.  The templates for use with this tool need to
       be updated to correct some details.  Gareth Seemungal is
       aware of the corrections needed."
   A.  Yes.
   Q.  Were you aware when you moved to Fujitsu what changes
       were being made to the templates for this tool or not?
   A.  No.
   Q.  If we look at the bottom box, "Fix impact", if you look
       at "Impact on user", well, let's look at "Impact on
       test", first:
           "Regression test ... with the new templates would be
       required and necessary as this tool directly inserts
       individual records into the branch database transaction
       tables ..."
           Yes?
   A.  Yes.
   MR JUSTICE FRASER:  What type of file is a .sh file?
   A.  Sorry, which line are we at?
   MR JUSTICE FRASER:  The one which says "Regression test
       of ..."?
   A.  Sh -- I think it is a shell script.
   MR GREEN:  "Impact on user:
           "Benefit of making the fix."
           If you go over {F/594/2}:
           "Corrects and updates the [transaction] correction
       tool templates, makes it less likely that mistakes will
       occur when fixing/resolving problems with transactions
       in BRDB."
           And then "Impact on operations":
           "Benefit of fix that may not be visible to end user.
           "SSC will be able to fix BRDB transactions quicker
       and with more confidence."
           Under "Risks":
           "SSC will have to make a note of the shortcomings of
       the existing templates and alter accordingly.
           "What are the risks of this fix having unexpected
       interactions with other areas?
           "Possible, if the templates are incorrect then
       branch totals, [transactions] ~... would be
       incorrect - hence the required regression testing."
           So that's the reason the regression testing is being
       undertaken, to try to avoid unintended consequences of
       the use of this tool.  That's right, isn't it?
   A.  Yes.
   Q.  What this shows us is that the continued use of the tool
       was at least anticipated at this time otherwise there's
       no point in updating the templates.
   A.  Agreed, yes.
   Q.  Can we look at {F/768} please.  This is PEAK number
       0208119 and you can see this is in 2012 and if we go --
   MR JUSTICE FRASER:  Well, the target date is 2012.
   MR GREEN:  Sorry, your Lordship is quite right.  The target
       date is 2012.  If we look at the progress narrative, it
       is created in February 2011, so again, spanning across
       the period that we have been looking at.
           Now, can we go down please just to get some context,
       can we go please to page 4 {F/768/4} and if we look at
       the second box, do you see the second paragraph there:
           "The SSC role has execute permissions to the
       Host-Dev delivered packages mentioned in the PEAK and
       resource, monitor (and hence connect and select any
       table) and select any dictionary."
           Now, pausing there, a select permission is not the
       same as an insert, delete or update permission, is it?
   A.  That's correct.
   Q.  Now, if we go up please to page {F/768/3} and look at
       the bottom of this, this is 30 September 2011:
           "As per the previous PEAK comments, the role
       'APPSUP' is extremely powerful and should only be used
       under extreme circumstances and under MSC supervision.
       As such the Branch Database design was that 3rd line
       support users should be given the 'SSC' role which is
       effectively read access ..."
           That's the select table and catalogue, yes?
   A.  Yes.
   Q.  "SSC team members should only have to ..."
           It looks as if it is access but it has got
       a little --
   A.  Yes.
   Q.  "... [access] BRSS for normal support investigations,
       unless the information has not replicated in time.  SSC
       should only given the optional 'APPSUP' temporarily ...
       if required to make emergency data amendments in BRDB
       live."
   A.  Correct.
   Q.  So pausing there, what was meant to happen was that SSC
       should only be given the optional APPSUP role
       temporarily, yes?
   A.  That's right, yes.
   Q.  And if we go up, please, to page 2 {F/768/2}, you can
       see Anne Chambers, halfway down the page -- do you see
       Anne Chambers?
   A.  I do.
   Q.  You know she has been involved in a lot of these PEAKs,
       hasn't she?
   A.  She has, certainly.
   Q.  And she is writing to Andrew Gibson; who is he?
   A.  Andrew Gibson is the guy in Ireland -- I can't remember
       his exact title, but he is certainly a key person in the
       operations area.
   Q.  And what she says is:
           "Unfortunately development write their scripts
       explicitly to use SSC.  So I think we're stuck with it
       unless they deliver new scripts (which would not be
       a popular or quick option).
           "When we go off piste we use appsup.  Can we have
       both??"
           Do you know from your knowledge what she meant by
       going off piste in that context?
   A.  It would be having to fix a problem that is not catered
       for by a script that is available to these people.
   Q.  Do you know that, or are you speculating?
   A.  I'm very confident in that.
   Q.  Finally in relation to these matters, you are aware that
       Ernst & Young in 2011 expressed concern about
       permissions -- privileged access permissions?
   A.  I am.
   Q.  And that was at the end of -- the Opus date on that
       document is 9 December 2011, yes?  And just pausing
       there, that is a very proper point for someone auditing
       these sorts of issues to be concerned about, isn't it?
   A.  Absolutely.
   Q.  I won't take you to it, but that letter is at {F/869}.
           If we now turn please to {H/196/9}, this is a part
       of a letter of 11 February 2019 and is from Wombles to
       Freeths and this user IDs point is being dealt with at
       paragraph 8.2 onwards:
           "We have taken further instructions from Fujitsu in
       it respect of additional USERIDs in your schedule and
       set out descriptions of each user below ... there are
       some usernames that: 1) do not exist today and never
       existed in the past as far as Fujitsu are aware.  These
       are identified in the classification column of the list
       as 'unknown', and 2) do not exist today but did exist in
       the past."
           So let's just take this in stages, if we may.  We
       can see COBEN01 is Cath Obeng but she also has
       a different entry below.  And she is "Person-SSC", yes?
   A.  Yes.
   Q.  And we can see a number of "Person SSC"s.  And then if
       we go halfway down we can see Andy Gibson is "Person-POA
       UNIX".  Do you see that?
   A.  Yes.
   Q.  And we can see that Ed Ashford is also "POA UNIX" on the
       next page, {H/196/11}, four down.
   A.  Yes.
   Q.  We can see at the bottom of the page Paul Simpson.  If
       we go over the page {H/196/12}, we can see Paul Stuart
       and Wayne Calvert, halfway down, "POA UNIX", and if we
       go please to page 14 of that letter, so two
       pages further on {H/196/14}, the explanation is given:
           "In respect of SSC user privileges, SSC users ..."
           That's "Person SSC":
           "... will typically have the following
       privileges ..."
           And then we can see what's said there.  For example
       at 8.6.3 a "Select" privilege.  But if we go to 8.7
       {H/196/15}:
           "In respect of POA Unix user privileges, POA Unix
       users (those identified as 'PERSON - POA UNIX' under
       classification) will typically have the following
       privileges/roles."
           And 8.7.7 is:
           "Granted the 'UNXADM' role which contains the
       following role:
           "(a) granted the 'DBA' role.  This is an Oracle
       supplied role for use by database administrators ...
       lots of privileges are granted to this role so users
       have the ability to update/delete/insert into any of the
       branch database tables."
   A.  Yes.
   Q.  You would agree that those people have the role which
       allows them privileges to update, delete, or insert into
       branch database tables whether they are using the
       correction tool or not?
   A.  Those people could log on to the database and do an
       awful lot of damage.
   Q.  And the only audit of that that we have prior to 2015
       was log on and log off; that's correct, isn't it?
   A.  Correct.
   MR GREEN:  My Lord, that's the end of that section of the
       cross-examination.  Would that be a convenient moment
       for a break?
   MR JUSTICE FRASER:  Yes.
   MR GREEN:  All I have left to deal with is Mr Godeseth is
       the bugs.
   MR JUSTICE FRASER:  Okay.  We will have ten minutes.
           Mr Godeseth, another short break.
   A.  Thank you, my Lord.
   MR JUSTICE FRASER:  If you could come back at 11.50, same as
       before, but you know the form: don't discuss the case
       with anyone.
   (11.41 pm)
                          (Short Break)
   (11.51 am)
   MR GREEN:  Mr Godeseth, I would now like to deal with the
       key bugs that you deal with in your second witness
       statement, if I may, at paragraph 3 at {E2/7/1} you
       explain that this witness statement is prepared in
       response to the witness evidence submitted by the
       claimants and the report submitted by Mr Coyne.  Is that
       right?
   A.  Yes.
   Q.  And at the end of paragraph 3 you say you have consulted
       with various colleagues and you have identified them
       below.
           Mr Bansal was involved in problem management; was he
       involved in these bugs, do you know?
   A.  I certainly haven't spoken to Steve about any of these
       bugs.
   Q.  You haven't spoken to him?
   A.  No.
   Q.  Let's look at Callendar Square, which you deal with from
       paragraph 12 onwards at {E2/7/3}.  And you say at
       paragraph -- you are dealing with this in the context of
       the trial of Seema Misra which you have referred to at
       paragraph 7, because you are responding to
       Mr McLachlan's witness statement as well, yes?
   A.  Yes.
   Q.  And at paragraph 8 you say:
           "I understand from Post Office's solicitors
       that ..."
           So some of this is just what you have been told by
       Post Office's solicitors?
   A.  Absolutely, yes.
   Q.  And at 8.3:
           "Post Office was represented [in that case] by
       Cartwright King ~..."
           Who informed Post Office's solicitors in turn about
       what you then say at 8.3, that's correct?
   A.  Yes.
   Q.  And you didn't really know about any of that yourself?
   A.  No.
   Q.  You have just effectively been asked to put it in your
       statement.  And at paragraph 10 {E2/7/3} you say you
       understand from Gareth Jenkins, halfway down:
           "... that Professor McLachlan requested information
       that was not necessary and/or obtainable."
           That's what Mr Jenkins told you about what
       Professor McLachlan had requested in that trial?  You
       didn't have any knowledge of it yourself?
   A.  The Misra trial -- I was in -- sorry, I was in Fujitsu
       so I have -- we would have had fleeting conversations
       about it but it was certainly not going down into any
       detail at that time, but I do remember the -- I do
       remember that we had been asked for an awful lot of
       information.
   Q.  But the characterisation --
   A.  And, you know, specifically the fact that people were
       wanting to keep the old Horizon system which just wasn't
       a viable option.
   MR JUSTICE FRASER:  What do you mean "keep"?
   A.  Basically to have a Horizon system up and running, an
       old Horizon system up and running, so that people could
       go and try to play with it to replicate errors or
       whatever.
   MR JUSTICE FRASER:  So you mean retain at least one part of
       it so that it would be available --
   A.  And --
   MR JUSTICE FRASER:  Will you just let me finish.
   A.  Sorry.
   MR JUSTICE FRASER:  When you say keep, you don't mean keep
       it for use across the network?
   A.  Certainly not, my Lord.
   MR JUSTICE FRASER:  You mean in respect of an operating
       version that was still available to be inspected?
   A.  Correct.
   MR JUSTICE FRASER:  Right.  And in your view, that wasn't
       viable?
   A.  You need to go beyond the counter to supporting
       components behind it in order to have anything that is
       likely to be of any use in trying to sort out how the
       whole system works, my Lord.
   MR JUSTICE FRASER:  Right.
   MR GREEN:  To your knowledge were there other requests made
       by Professor McLachlan for information, or is that not
       something you know about?
   A.  I don't know about the detail of that certainly.
   Q.  But what you do know you've got from Gareth Jenkins
       there, that Professor McLachlan requested information
       that was not necessary and/or obtainable, that's what
       you've got from Gareth Jenkins?
   A.  Yes.
   Q.  If we go to paragraph 12 onwards, you are specifically
       dealing with Callendar Square, yes?
   A.  Yes.
   Q.  And you very fairly say that you don't have first-hand
       knowledge of the issue, yes?
   A.  Correct.
   Q.  But you say you have reviewed the documents relating to
       the issue and discussed it with Gareth Jenkins.
   A.  Yes.
   Q.  So you have both undertaken a review of the documents
       and spoken to Mr Jenkins, that's correct?
   A.  Yes.
   Q.  That's what you say.
   A.  Yes, yes, I have certainly spoken to Gareth, but -- and
       I have re-clarified a couple of things just to make sure
       that my understanding was correct.
   Q.  Let's just start with Callendar Square.  Just very
       briefly, if I can take these reasonably quickly,
       firstly, it is a Legacy Horizon problem not
       a Horizon Online problem?
   A.  Correct.
   Q.  Secondly, it was a problem with Riposte?
   A.  Correct.
   Q.  Thirdly, it is sometimes referred to as the Riposte lock
       or unlock problem?
   A.  Yes, I'm happy with that, yes.
   Q.  And, fourthly, it affected balancing in branches?
   A.  It could affect balancing in branches.
   Q.  And let's start, if we may please, at {F/243}.  This is
       the KEL ending in 1359R and we can see that the KEL was
       raised on 24 November 2000.  Do you see under the
       "Raised" --
   A.  Yes.
   Q.  And you see it says "Time out waiting for lock agents
       and correspondence"?
   A.  Yes.
   Q.  If you look under "Solution - ATOS":
           "Needs investigating ... if we get an occasional one
       of these events, when Riposte is busy (such as Riposte
       starting up which was the occurrence on the 04/04/2003)
       then this can be ignored.  The original error type that
       this call was raised for was when the Riposte service
       hung following archiving and continuously output these
       messages."
           Yes?
   A.  Yes.
   Q.  This is a type of problem that could arise with Riposte
       lock agents, yes?  And on the face of --
   MR JUSTICE FRASER:  You are moving on before you got
       an answer.
   A.  Sorry, my Lord, yes.
   MR GREEN:  I thought I heard him say yes.
           And on the face of this, what's being said is if we
       get an occasional one of these events, this can be
       ignored.
   A.  Yes, an occasional time out when one thing is trying to
       talk to another is something that you could say is okay,
       but that's a highly subjective thing because you have to
       work out what "occasional" means.
   Q.  Okay.  Let's go forward to {F/565} please and look at
       page 2 {F/565/2}.  And if we look at February 2003, this
       is the J Simpkins 338Q KEL.
   A.  Yes.
   Q.  And if we look at February 2003:
           "We are seeing a few of these each week, on
       Wednesdays during balancing.  This can lead to problems
       if the PM is balancing on the counter generating the
       events, as it may not have a full view of transactions
       done on other counters.  PC0086212 sent to development."
           And then if we look at June 2004 further -- this is
       a year and a bit later on:
           "This event can also give rise to Transfer problems,
       where the eventing Node was not replicating and so
       allowed Clerk to Transfer in ... of a [transaction]
       which had already been TI [transferred in] on another
       Node for the second time or an Existing Reversal ..."
           Now, do you know if on the occasion being referred
       to there, there was a discrepancy in the branch or not?
   A.  I don't know, my Lord.
   Q.  It's not apparent from this KEL anyway.  Do you want to
       go over the page {F/565/3}:
           "This problem is still occurring every week, in one
       case at the same site on 2 consecutive weeks."
           It doesn't actually tell us whether there is
       a discrepancy at the branch, does it?
   A.  No.
   Q.  And this final entry of September 2005 is actually the
       Callendar Square branch so we can see that although
       Callendar Square experienced this bug in 2005, the
       genesis of it appears to go back to February 2003 and
       similar lock agent problems back in November 2000, is
       that fair?
   A.  I think that's fair.
   Q.  And if we go forward, please, to another KEL at {F/354}.
       This is raised in June 2004:
           "Transfer in accepted Twice, thus causing
       a Discrepancy".
           Now, we saw in June 2004 that there was a problem of
       a duplicate transfer in on the other previous KEL,
       didn't we, and this seems to be the June 2004 branch
       referred to in the previous KEL.  It says under
       "Problem", discrepancy of £22,290.  Do you see that
       figure about five lines down?
   A.  Yes.
   Q.  So that's the discrepancy.  Then if you look under
       "Solution", it is:
           "~... currently with [Escher] for investigation."
   A.  Yes.
   Q.  Yes?  At the end of that solution.  From this KEL we
       don't know if this problem was sent to Escher in 2004
       when it was first raised, or some time thereafter,
       or October 2006 when this KEL was last updated, do we?
   A.  I don't know, I don't know when it was actually sent to
       Escher.
   Q.  You can't tell from the KEL?
   A.  No.
   Q.  And then if we look at the Callendar Square branch
       itself more closely, {F/297}, this is PEAK 0126042 and
       you can see the Callendar Square branch FAD code of
       160868.
   A.  Yes.
   Q.  And the problem is in the summary, just after the FAD
       code:
           "SU cash amounts vary on counters."
           Yes?
   A.  Yes.
   Q.  And if we look -- we can see first of all
       15 September 2005, there's a call from Alan Brown.  If
       we go over the page {F/297/2} and it is the fifth box
       I think, 15 September 2005 at 16.12.27.  Do you see
       there:
           "Due to the Riposte errors on 14/09/05 from 15.30
       onwards ... messages were not replicated on counter 3.
       As a result, 3 transfers in to stock unit AA were done
       twice, initially on other counters then again on counter
       3.  The transfers in were for ..."
           And we can see the figures there:
           "This has resulted in a loss of 3,489.69 in CAP 25
       to the outlet, which POL may need to correct via an
       error notice."
           Do you see that?
   A.  Yes.
   Q.  And then underneath that:
           "Phoned the PM to explain what the problem was.  He
       is concerned about other transactions which he has input
       twice (3 Giro deposits and another cheque) because of
       the replication problem.  Have advised him to contact
       the NBSC as this is a business issue."
           Now, it's clear that there is an underlying -- well,
       we know the Callendar Square bug is a bug, don't we?
   A.  Yes.
   Q.  And if we move forward please to {F/298}.  This is PEAK
       0126376, and we can see it's the same FAD code, 160868,
       and over the page on page 2 {F/298/2} he has called back
       on 21 September 2005 at 17.09 and if you look under --
   MR JUSTICE FRASER:  21 September at when?  17.09?
   MR GREEN:  At 17.09, just the top of the page.
   MR JUSTICE FRASER:  Yes.
   MR GREEN:  "Information: pm experienced similar problems
       last week and wants us to investigate why this is
       happening.
           "PM stated he has not had any disconnected node
       messages or any online issues."
           Do you see that?
   A.  Yes.
   Q.  And immediately below that, information at 5.14:
           "Another transfer of £45.40 approx."
           And then if we go down further can you see:
           "This transfer appeared on both nodes 3 & 4."
   A.  Yes -- sorry, no.
   Q.  You can see "Transfer out - node 2", and then we go
       below that to "Transfer in node 3 & 4"?
   A.  Yes.
   Q.  And then that tells us -- that's the information to
       which the observation "This transfer appeared on both
       nodes 3 & 4", is made?
   A.  Yes.
   Q.  And:
           "When clerk accepted transfer in it accepted it.
       After she had accepted it she went back into transfers
       and node 4 transaction for £608.13 appeared."
           So I think this is actually an assistant at the
       branch, but it's the same branch.
   A.  Yes.
   Q.  And if we look on page 3 please {F/298/3},
       22 September 2005, 16.13.08, which is about four down
       from the top -- five down from the top.
   A.  Yes.
   Q.  Cheryl Card:
           "This is another occurrence of last week's
       problem ..."
           With a PEAK reference there:
           "... where messages were not replicated due to
       Riposte errors.  Last week the PM did some duplicate
       transfer in transactions which caused reconciliation
       errors - fortunately this week he has not done so.  The
       PM wants to know why these problems have occurred twice
       within 2 weeks.  I will pass the call to development for
       comment."
           Yes?
   A.  Yes.
   Q.  And you can see there that the -- it is a pending
       incident, yes?  Category 40, "Under investigation"?
   A.  Yes.
   Q.  If we move forward, please, to {F/300.1}, this is now
       from the SPM's point of view what's happening and he is
       actually getting some support on this particular one
       from his manager apparently here.  There are visits
       noted at 20th, 21st and 28th September 2005 and if we
       look at page 2, please, {F/300.1/2} you can see under
       "Expand on any letter requested ..." can you see:
           "This office had severe problems balancing on [week]
       25, resulting in a shortage of £6,414.46.  After
       checking various reports I am satisfied that the error
       is made up of ..."
           And they are then broken down carefully there and
       you can see that the SPMR is saying there is a Horizon
       software problem and so forth, giving the history, also
       a problem with transfers from one stock to another, they
       have doubled up.  Then speaking to the Horizon support
       centre at the bottom of the page.  If we go over the
       page please {F/300.1/3}:
           "They told the SPMR that they would report to NBSC
       that they had identified and rectified the problem and
       that the amount could be held in the suspense account.
       However, as part of the shortage relates to transfers,
       and no error notice will be issued, then the
       Suspense Account Team are not prepared to authorise the
       entry."
           So it wasn't easy for the subpostmaster, was it, you
       can see here?
   A.  No, indeed, a horrible position to be in.
   Q.  And we can see the concern about follow up if you look
       at the penultimate paragraph:
           "No follow-up was received from Service Support
       regarding this call."
           Now that's the Horizon service desk, isn't it?
   A.  It would not be Fujitsu.  Service Support, I think,
       would be a Post Office business function.
   Q.  Was it?  Okay, thank you.
           Let's move forward please to {F/301.2}.  This is
       another log, "Area Intervention Manager Visit Log",
       dated 7 October 2005, and if we go over the page
       {F/301.2/2}:
           "Telephoned Alan as requested.  He is concerned that
       he has still not heard anything regarding the loss that
       he is rolling."
           Yes?
   A.  Yes.
   Q.  Just tracing through the chronology.  If we go to
       {F/302} please, this is five days later,
       12 October 2005, and if you look on this first page you
       will see at 17.26, which is just below the second lot of
       equal signs:
           "PM is transferring a stock unit from node 7 to node
       2, it has been accepted on node 2 but has not been
       transferred."
           Do you see that?
   A.  I do.
   Q.  And then at 17.39:
           "PM was trying to transfer £2,490 from node 7 onto
       node 2.  She states that she has accepted the transfer
       on node 2 but the system is not showing this.
           "On node 7 it is showing pending transfer but it is
       not showing on node 2.
           "It appears on her transfer sheet as completed."
           So that would be very anxious-making for the SPM,
       wouldn't it?
   A.  Agreed, yes.
   Q.  If we go to {F/310.1/2}, 4 November 2005, page 2 please.
           "Spoke to Alan about the ongoing issue of his
       Horizon problems.  He now has received the error notice
       for the Girobank duplicate transactions.  This still
       leaves a loss of around £3,500.  I told Alan that this
       was still being investigated and as soon as I had some
       news I would get back to him.
           "Alan worried about this, because he has read an
       article in the Subpostmaster which makes him think that
       he will be held liable, and he has no intention of
       making this good.  He again said that the transfers had
       been sorted, but he needed paperwork to back this up in
       order to correct the loss."
           Yes?  Now, if we go back to the PEAK to see how this
       is being dealt with at {F/298/5} please.  We are now
       at -- this is part of PEAK 0126376, and if we look at
       10 November 2005 at 11.58.30, user Jez Murray, do you
       see there:
           "1.  This problem is the root cause of the
       reconciliation error closed in [previous PEAK].
           "2.  Presumably the root cause is deemed to be
       software not hardware.
           "3.  The Postmaster has a workaround in place which
       is not to duplicate transactions (eg Transfer In) just
       because the original attempts were successful but not
       showing on all nodes."
           And then 4, you can see:
           "POA CS MSU have a workaround in place which is that
       if 3 above is not followed and [it] recurs, a BIMS will
       be issued advising POL to issue a Transaction
       Correction."
           Then 5:
           "There is no SLT for software fixes as they are
       delivered based upon the priority or severity of the
       issue and could remain open until both businesses decide
       a fix is necessary or the workaround is adequate."
           So just unpacking paragraph 5 for a moment, there's
       no service level time for a software fix of this type,
       is that right?
   A.  I'm -- it says it there so I'm -- yes, I would -- I have
       no reason to doubt it.
   Q.  Sorry, service level target I should have said.
           Now, if we move forward, please, from that being
       left as a workaround to {F/312.1}.  This is another log
       of 167 November 2005.  If we look at page 2 of that one
       please {F/312.1/2} you can see there are two things
       referred to there:
           "Discussed ongoing issues of loss due to Horizon
       misbalance.  SPMR concerned that he has now made
       a fraudulent entry in that he has rolled over to the
       next trading period and put the loss into local
       suspense.  He has then gone on to state that the cash
       has been made good, which it hasn't.  This was done on
       the advice of the Helpdesk."
           Now, when you were at Post Office were you aware of
       that sort of advice being given to people who were
       suffering software or balancing problems?
   A.  No.
   Q.  18 November 2005:
           "I have contacted Jennifer Robson, who has confirmed
       that the loss has to be put into Emergency Suspense.
       I have now asked when this will be done as the SPMR is
       worried about doing something that is totally contrary
       to his contract."
           You can understand those worries, can't you?
   A.  I certainly can.
   Q.  Now, did any of the Callendar Square matters really
       bubble up to you, any of the aspects of this at the
       time, or not?
   A.  I think I was probably aware of Callendar Square as --
       but that would have been 2010 when I went into Fujitsu,
       so I cannot remember ever having heard of
       Callendar Square or Falkirk --
   Q.  Post Office.
   A.  -- Falkirk Post Office as an issue before that.
   Q.  Okay.  Let's look at {F/324.1} please.  We have 5 and
       6 January, it's another log:
           "Telephoned the office and Allan said that he was
       having problems again with transfers.  He has contacted
       the Horizon helpdesk who have subsequently come back to
       him to say that there is no system problem and that he
       should contact NBSC."
           Now, the Horizon help desk, was that one run by
       Fujitsu or Post Office?
   A.  My recollection is that the Fujitsu help desk would have
       been dealing with hardware issues and the NBSC would
       handle the process and business queries, so I think
       there was a Fujitsu help desk which was where the
       subpostmaster would ring to say he thought he had
       a problem with his hardware.
   Q.  Didn't the Fujitsu help desk also deal with queries
       about software problems?
   A.  Not talking to the Post Office -- not talking to the
       subpostmaster, as far as I can recall.
   Q.  Okay:
           "He did this and from what I can understand the NBSC
       have told him that he is trying to balance on two
       different terminals.  Allan disputes this and is adamant
       that there is a system error."
           Do you see that?
   A.  Yes.
   Q.  Now, given the history, it's quite likely that there was
       a system error, isn't it?
   A.  I think it's -- it has been established that there was
       a bug.
   Q.  Yes.  If we then move forward please to {F/333.1} and we
       go to page 3 first, please {F/333.1/3}.  This is part of
       an email chain and if we look at the email of
       23 February 2006 from it Anne Chambers at Fujitsu to
       Mike Stewart, you see that halfway down, Anne Chambers
       to Mike Stewart about Callendar Square.
   A.  Yes.
   Q.  If you look, please, at the second paragraph:
           "Haven't looked at the recent evidence, but I know
       in the past this site had hit this Riposte lock problem
       2 or 3 times within a few weeks.  This problem has been
       around for years ..."
           That's consistent with the related KELs we have
       already looked at, isn't it?
   A.  Yes.
   Q.  "... and affects a number of sites most weeks, and
       finally Escher say they have done something about it.
       I am interested in whether they really have fixed it
       which is why I left the call open - to remind me to
       check over the whole estate once S90 is live - call me
       cynical but I do not just accept a 3rd party's word that
       they have fixed something!."
           Yes?  So it had been open, we know, for years as
       a known problem affecting sites for years, and Escher
       are hopefully fixing it, although Anne Chambers is
       doubtful about that, in February 2006?
   A.  She is cynical, yes.
   Q.  Yes?  She goes:
           "What I never got to the bottom of, having usually
       had more pressing things to do, was why this outlet was
       particularly prone to the problem."
           So that doesn't suggest that she gave the problem of
       this outlet a particularly special priority to sort out
       its problems, does it, because she had more pressing
       things than an outlet being affected by significant
       discrepancies from a known bug?
   A.  Yes.
   Q.  "Possibly because they follow some particular
       procedure/sequence which makes it more likely to happen?
       This could still be worth investigating, especially if
       they have continuing problems, but I don't think it is
       worthwhile until we knew the S90 position."
           So she doesn't seem to be going to investigate it at
       that stage.  And then she says:
           "Please note that KELs tell SMC that they must
       contact sites and warn them of balancing problems if
       they notice the event storms caused by the held lock,
       and advise them to reboot the affected counter before
       continuing with the balance.  Unfortunately in practice
       it seems to take SMC several hours to notice these
       storms by which time the damage may have been done."
           Now, this isn't something you knew about at the
       time, but from what she is saying there it doesn't seem
       that the SMC information "watch out" is getting to the
       branches in time, is that fair?
   A.  That's a fair assumption, yes.
   Q.  And you say in your witness statement that release S90
       was distributed in March 2006.  That's paragraph 14.2 at
       {E2/7/5}.
   A.  Yes.
   Q.  So stepping back, we have had related problems since --
       directly related at least since February 2003 and other
       related ones we saw in the earlier KEL from 2000, and it
       takes them until 2006 to distribute the fix that deals
       with it, yes?
   A.  Yes.
   Q.  In circumstances where many branches had been affected,
       including their branch accounts.
           Now, can I ask you to look at {H/2/96} please.  This
       is part of a letter sent by Post Office's solicitors on
       28 July 2016.  If we look, this is dealing with
       Callendar Square, paragraph 2.1:
           "This defect, which was discovered in 2005 and fixed
       in March 2006, involved Horizon failing to recognise
       transfers between different stock units."
           Now, we have seen that it's not correct that the
       actual underlying defect was only discovered in 2005,
       haven't we?
   A.  I would agree that the bug underlying it -- the
       Callendar Square manifestation was discovered in 2005
       but --
   Q.  The one at Callendar Square?
   A.  Yes, and I would agree that the underlying bug had been
       there for a considerable time, probably since the
       Horizon went in.
   Q.  And if we look at the following -- sorry, if we look at
       {H/6/3}, this is part of a letter from Post Office's
       solicitors on 11 January 2017 and if we can look at
       page 3 of that letter very kindly, if you look at
       paragraph 7.8.1:
           "The Falkirk/Callendar Square issue was only known
       to have affected that one branch."
           That's also incorrect, isn't it?
   A.  Certainly there is a spreadsheet which actually talks
       about where we found this lock problem and how many had
       had a -- had discrepancies in there -- how many we
       thought would have had -- there would have been an
       impact of this type.
   Q.  That's wrong, isn't it?
   A.  On the face of it, yes.
   MR JUSTICE FRASER:  Well, you might want to look also at --
       it is a PEAK that Mr Green took you to, but he didn't
       draw your attention to the entry that I'm going to draw
       your attention to, just out of fairness.  It is at
       {F/298/3} -- actually it starts at {F/298/1} so you can
       see it is one of the PEAKs that Mr Green has been asking
       you about and he took you to different entries, but if
       we go now to page 3 of that {F/298/3} you will see in
       the bottom half of the page in the light blue, entry
       22 September 2005 from Cheryl Card, and in parentheses
       at the bottom of that entry she says:
           "A few of these errors seem to occur every week at
       different sites."
           Does that help you answer Mr Green's question?
   A.  I think that -- the "timeout occurred waiting for lock",
       so this is a generic -- I would read that as a fairly
       generic problem which is probably the one which goes
       right back to the ATOS KEL -- sorry, the KEL which says
       that if you see this occasionally, that might be okay,
       but it's really when you get a storm of them that you
       have effectively got one counter not being able to talk
       to another and that is the underlying technical issue is
       that a counter is simply not being able to get a message
       across to another one.
   MR JUSTICE FRASER:  Mr Green.
   MR GREEN:  And I think you fairly accept that there is
       a spreadsheet, which we will come to, with a significant
       number of branches that were affected by the
       Falkirk/Callendar Square bug.
   A.  Yes.
   Q.  And that spreadsheet includes more than one entry,
       doesn't it?
   A.  It does.
   Q.  Let's look at your witness statement, please, at
       {E2/7/3}.  You say the Callendar Square bug occurred in
       2005.  It would be more complete to cover the history
       that we have covered now, wouldn't it?
   A.  That's a fair comment, yes.
   Q.  At paragraph 15 {E2/7/5} you say:
           "I understand from Matthew Lenton (Fujitsu's
       Post Office Account Document Manager) that this bug
       affected thirty branches, resulting in mismatches at
       twenty, and that Fujitsu has established this for the
       purposes of this statement using the event logs
       described above."
           Is that really what Matthew Lenton told you?
   A.  I didn't speak to Matthew about it.  I got ... so the
       answer is no, that's not what Matthew told me.
   Q.  Did Matthew communicate directly with you in any way on
       this point?
   A.  I think I got the spreadsheet from him.
   Q.  Directly from him, or from someone else?
   A.  I can't remember.
   MR JUSTICE FRASER:  I think to be fair, you actually
       corrected that number in-chief, didn't you?
   A.  I did indeed, my Lord.
   MR GREEN:  Yes, I will come to that.
           And just to clarify, is it your understanding from
       the information that's found its way to you from
       Matthew Lenton that there has been a specific process
       for the purposes of your statement of analysing the
       event logs historically across all the branches that may
       have been affected by this to see how many were?
   A.  No, the spreadsheet is something that was done at the
       time, as I understand it by Anne Chambers, to actually
       investigate what had happened when we had had these
       locks.
   Q.  Because the statement there would suggest that we can be
       completely confident that Fujitsu have established
       what's happened because they have looked at all the
       event logs, and that's not right, is it?
   A.  I am comfortable that Anne will have done her job
       correctly and she will have investigated all of those
       issues at the time, resulting in the spreadsheet which
       I believe we found in Steve Parker's -- he had -- he
       still had it in it his audit trail.
   Q.  I will just put the question one more time: the
       statement that's made there is not right, is it?
   A.  I agree, it's not right.
   Q.  Now let's look at {F/322.1} please.  We can download
       that.  This appears to be the statement that you have
       referred to in oral evidence now but that was not
       referred to in your witness statement and we will see in
       a minute that the properties identify Anne Chambers as
       the author on 22 December 2015 and --
   MR JUSTICE FRASER:  Maybe just pause while we wait for it to
       come up.
   MR GREEN:  Could we just make it a little bit bigger, very
       kindly.
           Now, this is the spreadsheet you are referring to?
   A.  Correct.
   Q.  That you have just told the court about.  We located
       this and it was disclosed to us on 27 February 2019 and
       if we look at the file tab at the top, please, and go
       into -- just pause there.  Can you see "Created", on the
       right-hand side, 22/12/2015 and "Anne Chambers", do you
       see that?
   A.  Yes.
   Q.  Is that how you identified it was from Anne Chambers, or
       did Mr Parker give you an email or tell you about that?
   A.  Steve -- yes, Steve and I discussed it so I -- I think
       I originally got the information that it was
       Anne Chambers' spreadsheet from John Simpkins when I was
       asking him for -- because I lost my copy of it so
       I asked him for another copy and he explained to me that
       it was from Anne Chambers originally.
   Q.  So you have always known that the actual source of the
       information that you're talking about in your witness
       statement was this Anne Chambers spreadsheet, haven't
       you, because you say you lost your copy of it?
   A.  Yes.  I certainly became aware of this spreadsheet and
       it was available to me.  I at first did not know it was
       Anne Chambers who had created it.
   Q.  Well, let's go to the content of it please.  If we look
       at the box at the top under D2 -- 2D:
           "NB many other branches had multiple events,
       preventing replication, but these are the majority of
       those which came to PEAK, having either reported
       a problem or it caused a reconciliation report entry.
           "From Sept 2005, cash accounts were replaced by
       branch trading statements and the TPSC256 report was no
       longer populated.  I can't remember how we then knew
       about receipts and payments mismatches and if we would
       have picked up on further issues."
           You said -- first of all, it's clear that this is
       not a document which mirrors or reflects the accuracy
       which one might have understood from your paragraph 15;
       that's fair, isn't it?
   A.  Fair comment.
   Q.  And secondly, you said it's a document she made at the
       time, but we can see it was created in 2015 from the
       properties.
   A.  She created it in 2015, yes.  Sorry, 2005.
   MR JUSTICE FRASER:  She created it in 2005?
   A.  Well, sorry, I can't remember the exact -- I'm accepting
       the date of creation that was shown to me earlier --
       sorry, it's 2015, sorry -- I had misinterpreted that
       entirely.  My understanding is that this was created at
       the time by Anne Chambers.  Sorry, I was misreading that
       and seeing it as 2005.
   Q.  Let's move on to the receipts and payments mismatch bug
       if we may now.  Let's look please at E2/7 --
   MR JUSTICE FRASER:  Just before you move off ...
           (Pause).
           Just one question.  Was it the number of branches
       shown on this document that led to you correcting the
       figure of 20 to 19 in your evidence-in-chief?
   A.  I personally didn't -- I think it was a -- I certainly
       did not correct that.  We realised that my statement was
       out of line with what this spreadsheet said, and so it
       was somebody else who noticed it.
   MR JUSTICE FRASER:  Understood, all right.
           Right, Mr Green, go on.
   MR GREEN:  Let's move to receipts and payments mismatch, if
       we may.  You would fairly accept, I think, that this is
       a shorthand for a number of different versions of a bug,
       or a number of different bugs that can be helpfully
       grouped together, is that fair?
   A.  I'm trying to get my head into the receipts and payments
       one.
   Q.  Well, it's -- if it helps, it seems to become
       a shorthand for a specific bug in Horizon Online which
       arose in 2010.
   A.  Horizon Online 2010, I think was local suspense.
   Q.  Well, let's just have a look.  Your paragraph 36
       {E2/7/10} is in the section that begins on the previous
       page at {E2/7/9}.  You are dealing with --
   A.  Sorry, yes, it was Horizon Online, yes.
   Q.  Yes.  Just to orientate you, let's go back to your
       paragraph 36 {E2/7/10} you say:
           "In September 2010 a bug in HNG-X (Horizon Online)
       caused a receipts and payments mismatch in some
       branches."
           Do you see that?
   A.  Yes.
   Q.  Then you say:
           "I understand from Gareth Jenkins that the issue was
       initially identified as a result of Fujitsu's monitoring
       of Horizon Online system events and not as a result of
       calls from branches.  Fujitsu's Software Support Centre
       ~... which provides 2nd and 3rd line support, monitors
       system events as a matter of course."
   A.  Correct, yes.
   Q.  So your evidence is that it arose in September 2010,
       yes?
   A.  Yes.
   Q.  And Fujitsu identified it in September 2010?
   A.  I have not checked the detail, but I'm comfortable with
       that, yes.
   Q.  Is this something that you actually know about yourself,
       or have you got all this from Gareth Jenkins, as seems
       to be the case from --
   A.  Gareth explained to me how the problem had occurred and
       also -- yes, we went through the logic of what was
       happening when a branch was going through this sequence
       and why there would be a receipts and payments mismatch.
   Q.  So is it something you actually know about yourself, or
       is it only -- is your knowledge derived really from what
       Gareth Jenkins has told you?
   A.  It is knowledge derived from what Gareth told me, plus
       further looking at some detail.
   Q.  And the document you repeatedly refer to, for example,
       at paragraph 39, 40 and 41, is 29 September 2010.  Let's
       look at that.  It's at {F/1000}.  It is a document
       written by Gareth Jenkins, do you see that?
   A.  Yes.
   Q.  And if we go forward, please, to the document at
       {F/1001}, this is the receipts and payments mismatch
       issue notes.
   A.  Yes.
   Q.  And we can actually see Mr Jenkins' name there, but not
       yours?
   A.  Correct.
   Q.  You weren't involved in this at the time, were you?
   A.  No.
   Q.  And you don't refer to that document, even though
       Mr Jenkins would have known about it?
   A.  Agreed.
   Q.  And at {F/777}, this is a receipts and payments mismatch
       document, Fujitsu document, and you can see the author
       is Mr Jenkins again.
   A.  Yes.
   Q.  That's not a document that you refer to in your witness
       statement either, and this time there's been a BBC
       documentary broadcast on Monday 7 February 2011, yes?
   A.  Yes.
   Q.  So I think that's separate to the later Panorama
       programme we get in 2013.
           Did Mr Jenkins draw your attention to either of
       those two documents that we have just seen to help you
       get a fuller picture, or not?
   A.  No.
   Q.  No.  It's clear who actually knows about this; it's not
       you, is it?
   A.  Gareth certainly understands the detail.
   Q.  If we can go back to {F/1001}, this appears in fact to
       be somewhere between 29 September 2010 and
       4 October 2010, this document.  Page 1 says in the
       second paragraph, underneath "What is the issue?" which
       explains that:
           "Discrepancies showing at the Horizon counter
       disappear when the branch follows certain ~... steps but
       will still show within the back end branch account ...
       impacting circa 40 branches since migration ... with an
       overall cash value of~... £20K loss."
           And:
           "This issue will only occur if a branch cancels the
       completion of the trading period, but within the same
       session continues to roll into a new balance period."
           Then you see:
           "At this time we have not communicated with branches
       affected and we do not believe they are exploiting this
       bug intentionally."
           Yes?
   A.  Yes.
   Q.  So this was a known problem, wasn't it?  And you can see
       Mr Winn of POL Finance is there and Fujitsu are there?
   A.  Yes.
   Q.  Are you surprised that this had not been communicated so
       people would just be warned about it?
   A.  There was obviously a fear that subpostmasters may be
       looking to exploit this because it gave -- there was
       a fear that people could see this as a way of defrauding
       the Post Office.
   Q.  So concealing it from SPMs who were honest was justified
       because of the expectation of dishonesty of
       subpostmasters in the network, in a nutshell?
   A.  In my view, this was a decision made by Post Office on
       how to manage this particular bug.  You could interpret
       it the way that you have put it.
   Q.  Do you agree with the way that I have put it to you?
   A.  I think I'm agnostic.  I can see the -- I can see
       a rationale for not broadcasting this, but equally, if
       the objective is to be totally open and honest and take
       the risk of causing more chaos in the network then yes,
       I would have to agree with that.
   Q.  Let's look at page 2, please {F/1001/2}.
   MR JUSTICE FRASER:  Have you seen this document before
       today?
   A.  I have.
   MR GREEN:  At the top of the page:
           "Note at this point nothing into feeds POLSAP and
       Credence ..."
           And then if you just look at the end, Credence --
           "~... so in effect the POLSAP and Credence shows the
       discrepancy whereas the Horizon system in the branch
       doesn't."
           So there's information in POLSAP and Credence that's
       not available to the SPM, isn't there?
   A.  Yes, that's clearly the case.
   Q.  And so the branch will then believe they have balanced;
       yes?  Let's look at page 3 please {F/1001/3}:
           "The Receipts And payment mismatch will result in an
       error code being generated which will allow Fujitsu to
       isolate branches affected this by this problem, although
       this is not seen by the branches.  We have asked Fujitsu
       why it has taken so long to react to and escalate an
       issue which began in May.  They will provide feedback in
       due course."
           So given that we are either at the very end
       of September or the beginning of October, this document
       which you said you have seen before clearly shows that
       it wasn't only in September 2010, was it?
   A.  No.
   Q.  And therefore the statement in your witness statement at
       paragraph 36 that in September 2010 a bug caused
       a receipts and payments mismatch in some branches is
       materially wrong, isn't it?
   A.  On the basis of what we have just been through I think
       I have to agree.
   Q.  And you can see that the roll-out is planned for
       21 October, yes?
   A.  Sorry ..?
   Q.  "~... full roll-out to the network completed by 21st of
       October"?
   A.  Sorry, where are we looking now?
   Q.  This is on page 3:
           "Fujitsu are writing a code fix which stop the
       discrepancy disappearing ..."
           Do you see?
   A.  Yes.
   Q.  "... from Horizon in the future.  They are aiming to
       deliver this into test week commencing 4th October.
       With live proving at the model office week commencing
       11th October.  With full roll-out to the network
       completed by the 21st of October.  We have explored
       moving this forward and this is the earliest it can be
       released into live."
           So just pausing there, it has taken since May,
       from May to October to be corrected and in the
       meanwhile, branches appear to be kept in the dark, at
       least up to the date of this document.
   A.  Yes.
   Q.  And then what is being considered is which of three
       solutions to apply to the impacted branches.  The
       group's recommendation at the time is that solution 2
       should be progressed.  Now, just pausing there, do you
       know which of these solutions was in fact progressed?
   A.  My understanding is that TCs were issued.
   Q.  Let's look at what the considerations were.  Option 1,
       solution 1 was:
           "Alter the Horizon Branch figure at the counter to
       show the discrepancy.  Fujitsu would have to manually
       write an entry value to the local branch account."
           Yes?
   A.  Yes.
   Q.  So this was being openly discussed between Fujitsu and
       Post Office, that it was possible to manually write an
       entry to the local branch account, wasn't it?
   A.  Clearly, yes.
   Q.  "Impact - when the branch comes to complete next Trading
       Period they would have a discrepancy, which they would
       have to bring to account."
   A.  Yes.
   Q.  "Risk - this has significant data integrity concerns and
       could lead to questions of 'tampering' with the branch
       system and could generate questions around how the
       discrepancy was caused.  This solution could have moral
       implications of Post Office changing branch data without
       informing the branch."
           So on any view, the moral implications of changing
       branch data without informing the branch were recognised
       there, weren't they?
   A.  Yes.
   Q.  Solution 2:
           "P&BA will journal values from the discrepancy
       account into the customer account and recover/refund via
       normal processes.  This will need to be supported by an
       approved POL communication.  Unlike the branch 'POLSAP'
       remains in balance albeit with an account
       (discrepancies) that should be cleared.
           "Impact - Post Office will be required to explain
       the reason for a debt recovery/refund even though there
       is no discrepancy at the branch.
           "Risk - could potentially highlight to branches that
       Horizon can lose data."
           That sounds defensive, doesn't it?
   A.  Agreed.
   Q.  Solution 3:
           "It is decided not to correct the data in the
       branches (ie Post Office would prefer to write off the
       'lost'.
           "Impact - Post Office must absorb~... 20K loss.
           "Risk - huge moral complications to the integrity of
       the business, as there are agents that were potentially
       due a cash gain on their system."
           Go over the page -- that appears to be the end of
       that solution.  Now, this does not suggest, does it,
       being open and transparent to SPMs about the problem
       with Horizon?
   A.  Again, if -- I can accept that I can also understand the
       reason to want to manage this carefully.
   Q.  And in fact, this problem became so well-known that in
       the end, Post Office did communicate with SPMs, are you
       aware of that?
   A.  No.  I -- I will take your word for it.
   Q.  At paragraph 42, at {E2/7/11}, you said 60 branches were
       affected and you said:
           "... we can be sure that all instances of the bug
       were identified ..."
           Yes?  That's now been corrected to 62, hasn't it?
   A.  Yes.
   Q.  And at paragraph 45 {E2/7/12} you refer to a spreadsheet
       of affected branches which we had to seek disclosure of.
       That's at {F/754}.  If we could just make that larger
       please.  If we look at the "Affected branches" tab
       first, there's a list of affected branches there, and if
       we go back to the "Check for duplicate branches",
       there's a total at the bottom of 64 but we can see the
       number 2 in row 34 and the number 2 in row 41, can't we?
   A.  Yes.
   Q.  So if you subtract both those from 64 you get the 60
       number in your original witness statement, don't we?  Is
       that a calculation that you made or someone else made?
   A.  No, no, I didn't do a calculation to come up with the
       60.  I was quoting from other people.
   MR JUSTICE FRASER:  Someone just gave you the 60, did they?
   A.  I thought I was quoting from other people.  I -- Gareth
       even said to me that in my statement I had said
       "approximately 60", so I was not -- clearly I didn't
       because the statement here doesn't contain that word.
       I had rather hoped it had when this was first brought to
       my attention, but no, I certainly did not do any
       specific calculation to come up with the 60 that I put
       into my original statement.
   MR GREEN:  Did Gareth explain the change to you from 60 to
       62?
   A.  No.
   Q.  But how did he come to tell you what was in your
       original statement?  What was that conversation?
   A.  I picked up the number -- to be -- my objective in this
       was to explain to the court the symptoms of the bug and
       how -- the technical aspect of it.  I did not pay
       particular attention to getting the detail on how many
       branches were affected, correct.
   Q.  Okay.  So you have spoken to Gareth since your
       statement?
   A.  I don't think I have spoken to him about this in
       particular.  I was -- as I say, when Gareth had said
       I had originally said "approximately 60" I was thinking
       that was quite neat, but that's not the case.
   Q.  Well, you said "Gareth even said to me that in my
       statement I had said 'approximately 60'", so he must
       have said that to you after your statement had been
       filed?
   A.  It was a comment in a document that we were exchanging.
   Q.  But you hadn't spoken to him about remote access since
       your first statement?
   A.  No.
   Q.  Why have you stayed off that topic with him?
   A.  Oh, sorry, this was just a comment.  We have been
       exchanging documents, we have been commenting on
       documents, so it was not a particular conversation.  It
       is merely a case of Gareth had commented on this when it
       was pushed back to us that I had originally said 60 and
       actually the answer was different.
   Q.  If we look please at the "Affected Branches" spreadsheet
       for a moment please and if you look at line 7, do you
       see there minus 777 in July 2010?
   MR JUSTICE FRASER:  I'm sorry, which line are we on?
   MR GREEN:  I'm so sorry, row 7.
   MR JUSTICE FRASER:  Row 7?
   MR GREEN:  16 July 2010.  That's long before Mr Jenkins has
       written any notes about this at the end
       of September 2010, isn't it?
   A.  The event certainly happened before Gareth's -- yes.
   Q.  And presumably at that time the branch has got an
       unexplained discrepancy in its accounts, yes?  Because
       that was the point?
   A.  The branch wasn't actually seeing the discrepancy,
       was it?
   Q.  There is a discrepancy in the branch's accounts at that
       time.
   A.  Sorry, yes, yes, there has to be, because there's a
       receipts and payments mismatch yes.
   Q.  At paragraph 43 of your witness statement which is at
       {E2/7/11}, you say that you understand transactions were
       not injected to resolve this problem.
   A.  Correct.
   Q.  Can we look at {H/2/25} please, paragraph 5.16.3 says --
       if you can see the bottom three lines of 5.16.3:
           "As far as Post Office is currently aware
       a balancing transaction has only been used once to
       correct a single branch's accounts (not being a branch
       operated by one of the Claimants)."
           And if we look at footnote 57 it says there:
           "This was in relation to one of the branches
       affected by the 'Payments Mismatch' error described in
       Schedule 6."
           Does that accord with your understanding?
   A.  I don't know the detail.
   Q.  If we look at {H/2/97}, "Payments mismatch" is referred
       to there.
   A.  Yes.
   Q.  But you don't know whether the balancing transaction was
       used in relation to that or not, is that your evidence?
   A.  I don't know.
   Q.  And then if we look please at {C4/2/15} -- my Lord,
       I will be two minutes, if that's all right --
       question 39:
           "Please explain the reason(s) why this Balancing
       Transaction was made."
           And if we go over the page please {C4/2/16},
       "Defendant's response 39":
           "The Subpostmaster affected is not a Claimant and
       the Claimants do not require the information sought in
       order to understand Post Office's generic case or to
       plead a Generic Reply.  Post Office repeats the General
       Response.  However, the Balancing Transaction was used
       to correct an error arising from the 'Payments Mismatch'
       problem (as to which, see page 25 of the Letter of
       Response)."
           And at page 24 of that underlying document it is
       signed with a statement of truth, so on the face of this
       it is suggesting that a transaction was injected to at
       least one account.  Are you aware of whether that's
       correct or not?
   A.  I would have to go and check the FAD code of the branch
       where I know we did a balancing transaction and compare
       it with the FAD codes that you're looking at here.
       I haven't done that.
   Q.  And if they weren't the same, it might appear there was
       more than one?
   A.  If they weren't the same, my interpretation would be
       that this statement is incorrect.  There has only been
       one balancing transaction.
   MR GREEN:  My Lord, I only have local suspense, briefly, and
       Dalmellington bug to go.  Is that a convenient moment to
       break?
   MR JUSTICE FRASER:  I think so.  We're going to come back at
       2 o'clock, Mr Godeseth, so same form as before, don't
       chat to anyone about the case.  If you could come back
       at 2 o'clock.
   (1.02 pm)
                    (The luncheon adjournment)
   (2.02 pm)
   MR JUSTICE FRASER:  I have received an application,
       Mr De Garr Robinson.  Do you know about it?
   MR DE GARR ROBINSON:  My Lord, I know that there has been an
       application, that is almost all I know.
   MR JUSTICE FRASER:  Has it been served on the claimants?
   MR DE GARR ROBINSON:  My understanding is it has, my Lord,
       yes.
   MR JUSTICE FRASER:  Do you know about it?
   MR GREEN:  My Lord, we have just seen it, yes.
   MR JUSTICE FRASER:  I just saw it five minutes ago.  It is
       an application for me to recuse myself as being the
       managing judge in these proceedings which means
       effectively -- and also for this trial to stop.
       Although the application says "adjourn the trial",
       I think it really means start it again with another
       judge.
   MR DE GARR ROBINSON:  I haven't seen the application,
       my Lord.
   MR JUSTICE FRASER:  No, no.  That's what the application is.
       So the practical effects of that are that application
       has to be dealt with as soon as possible.
   MR DE GARR ROBINSON:  Yes.
   MR JUSTICE FRASER:  Simply in terms of timetabling, it
       presents some difficulties because we're in the middle
       of cross-examining a witness, so the options are for
       this trial to stop now so that application can be heard
       on a day which we will then come on to discuss, or we
       finish with Mr Godeseth and then the trial comes to
       an end, or is, rather, subject to an interval depending
       on what the outcome of the application is, so do you
       have any submissions about which of those two courses of
       action --
   MR DE GARR ROBINSON:  My Lord, I didn't appreciate that
       those were the options that you would be presenting now.
       I do -- entirely without instructions I would say it
       would be invidious for Mr Godeseth to remain in the
       witness box, as it were, for an extended period of time.
   MR JUSTICE FRASER:  Well, in principle I agree, but that's
       dependent on a number of factors, one of which is how
       much longer his cross-examination is going to be.
           Mr Green, how long do you think you're going to be?
   MR GREEN:  I think about 10 to 15 minutes.
   MR JUSTICE FRASER:  And how long would your re-examination
       be?
   MR DE GARR ROBINSON:  10 to 15 minutes.
   MR JUSTICE FRASER:  Right, well, this is my proposal.  I'm
       going to rise now for ten minutes, so you can take
       instructions.  If you are minded to persuade me to stop
       instantly then you can tell me at 2.15 and I will make
       a ruling on that.  If you are, in light of the
       application, content for Mr Godeseth's evidence to
       finish, in other words, it will be brought to an end
       with the cross-examination finishing, re-examination
       finishing, and any questions from me being dealt with,
       we will do that, and then we will address other
       logistical issues immediately after that.
   MR DE GARR ROBINSON:  My Lord, yes.
   MR GREEN:  My Lord, I'm sorry to interrupt, would
       your Lordship be amenable if the applicant in the
       application agreed to also hearing Mr Parker this
       afternoon so that --
   MR JUSTICE FRASER:  Well, that depends, because at the
       moment the Post Office hasn't necessarily asked for
       Mr Godeseth necessarily to be finished.
   MR GREEN:  My Lord, of course, but I mean --
   MR JUSTICE FRASER:  On the basis that if the application is
       successful this trial will have to be stopped and
       restarted, I'm not going to pre-judge the outcome of
       that application, but if Mr De Garr Robinson wants to
       take instructions on that then he can do that at the
       same time.
   MR GREEN:  Of course.
   MR JUSTICE FRASER:  All right, so is until 2.15 going to
       give you long enough?
   MR DE GARR ROBINSON:  My Lord, I would have thought so, yes.
   MR JUSTICE FRASER:  All right.  If you want any more time
       then just send a message.  I will come back in at 2.15.
           I'm afraid this is going to continue a little
       longer, but you've got until 2.15, Mr Godeseth, so if
       you can come back to the witness box for 2.15.
   (2.05 pm)
                          (Short Break)

   (2.18 pm)
   MR DE GARR ROBINSON:  My Lord, we have not been able to make
       contact with counsel instructed in the application.
   MR JUSTICE FRASER:  Are you not instructed?
   MR DE GARR ROBINSON:  I'm not instructed in the application
       at all.  My Lord, it is Lord Grabiner from my chambers
       who will be acting.
   MR JUSTICE FRASER:  Yes.
   MR DE GARR ROBINSON:  I have of course taken instructions.
       The consensus on this side of the court is that without
       prejudice to that application, it would be invidious for
       Mr Godeseth to be left in limbo for an unidentified and
       indeterminate period of time, so I suspect that it will
       be common ground on all sides that in fairness to him,
       his evidence should be finished.
   MR JUSTICE FRASER:  Well, as long as that's -- when I said
       take instructions, as long as I mean that that's
       instructions you have taken from your solicitors who are
       in court, then I'm completely content.
   MR DE GARR ROBINSON:  Yes.
   MR JUSTICE FRASER:  All right.
   MR DE GARR ROBINSON:  My Lord, as regards Mr Parker,
       pragmatically we are all here ready to ask him
       questions.  On the other hand, your Lordship may not
       feel comfortable dealing with Mr Parker's evidence given
       the current state of affairs, and I'm happy to -- I have
       no application that you should hear his evidence,
       your Lordship won't be surprised to hear me say that.
   MR JUSTICE FRASER:  No, but you are effectively neutral, are
       you?
   MR DE GARR ROBINSON:  My Lord, I don't want to press upon
       your Lordship continuing with his evidence.  I know
       my learned friend is anxious that he doesn't -- he needs
       half a day with Mr Parker and he is anxious not to
       lose -- we have already lost 15 minutes.  But if
       your Lordship were minded to continue with Mr Parker,
       given the pragmatic situation we are in, I wouldn't
       oppose that.  Entirely without prejudice, of course, to
       the application.
   MR JUSTICE FRASER:  All of it is obviously without prejudice
       to the application.
   MR DE GARR ROBINSON:  Yes, but I keep saying it, don't I.
   MR JUSTICE FRASER:  I was not asking you the question in any
       way to either dilute or change whatever the
       Post Office's position is on the substantive
       application, it was just a matter of trial management.
       If we do continue and deal with Mr Parker as well, that
       would at least have the advantage that it would
       effectively be a clean break at the end of the evidence
       of fact.
   MR GREEN:  My Lord, there is still Mr Membery.
   MR JUSTICE FRASER:  I haven't forgotten about Mr Membery,
       but he is not coming on today I don't think, is he?
   MR GREEN:  Not today, my Lord.
   MR JUSTICE FRASER:  I will revisit this at the conclusion --
       I'm completely comfortable doing it, but I will revisit
       it at the conclusion of Mr Godeseth because apart from
       anything else, we don't know how long that is going to
       take.
   MR DE GARR ROBINSON:  My Lord, yes.
   MR JUSTICE FRASER:  All right, Mr Green.
   MR GREEN:  Mr Godeseth, can we look at paragraph 46 of your
       witness statement at {E2/7/12} please, and this bug
       caused entries from local suspense accounts that had
       been there in 2010 to be reproduced in two successive
       years, yes?
   A.  Yes, that's right.
   Q.  And so this was something that persisted for years not
       months as a bug, wasn't it?
   A.  There was data in the system that shouldn't have been
       there for a couple of years.
   Q.  And if we look, please, at {F/1075}, this is a document
       that you have referred to in your witness statement.
       Now, its author is Gareth Jenkins, isn't it?
   A.  It is.
   Q.  And you have referred to understanding various matters
       from Mr Jenkins at paragraphs 48 and 54 of your witness
       statement.  Is it fair to say that you've got most of
       your information about this issue from Mr Jenkins?
   A.  Yes.
   Q.  And if we look over the page at page 2, please,
       {F/1075/2} do you see "The Problem" at 2.2?
   A.  I do.
   Q.  And if you look at the screen and I read out the first
       line of your 48.1:
           "Data in the Branch Database is regularly
       archived/deleted when it is no longer required."
           There's a difference there because a slash has been
       put in instead of "and then", yes?  Otherwise the
       sentence is identical.  Then it says:
           "There are different rules for how long different
       types of data is retained and when it is deleted."
           That is substantially the same, but with some
       slightly different words, and we see other features
       appearing, for example, at paragraph 50 of your witness
       statement, at the bottom of that page {E2/7/13}.
           This document has been substantially the source of
       that part of your witness statement, hasn't it?
   A.  Correct.
   Q.  And if we look at paragraph 48.2 of your witness
       statement {E2/7/13}, it says:
           "As a result of some changes that were made on
       3 July 2011 to the Horizon archiving strategy relating
       to Stock Units that had been deleted in a branch, any
       branch that deleted a Stock Unit at the end of 2010
       which had a local suspense transaction in that Stock
       Unit before it was deleted were left in the table used
       for constructing the branch trading statement."
           Now just pausing there, that's essentially how the
       local suspense bug operated, isn't it?
   A.  The local suspense bug came about because there were
       records in the branch database which should have been
       removed -- whether you call that archiving or deleting,
       they should not have been encountered by this particular
       bit of processing.
   Q.  And this is a good example of how small changes to what
       you might term non-transactional data can cause effects
       to branch accounts, is that fair?
   A.  The record here was very much used in constructing the
       branch trading statement.
   Q.  So does that suggest an interface between -- there's
       a sort of -- it's not a clean break between operational
       and transactional data, is it?  There's a sort of
       interlacing of the two where data has more than one
       purpose.
   A.  The purpose of this particular record was really to keep
       a running total of something, because you cannot go back
       to the complete raw data in order to come up with the
       branch trading statement, that would be a nightmare, and
       so the process that it goes through is that data gets
       summarised up into totals on a daily basis and you also
       build up things such as suspense values and -- over
       a period and then when you are producing the branch
       trading statement, these things all get brought
       together.
   Q.  Can I just ask you to look at paragraph 47 of your
       witness statement please on page 12 {E2/7/12}:
           "The bug was discovered in January 2013 when two
       Subpostmasters, who experienced the largest
       discrepancies, raised the issue with Post Office.
       I understand from Gareth Jenkins that Post Office could
       see the impact of the problem in their back end system
       and wrote off the discrepancies~..."
           The point there is that although it had persisted
       for years not months, it was SPMs who drew it to
       Post Office's attention and when they did, then Fujitsu
       saw it and was able to try and correct it?
   A.  We became aware of it in 2013, from my understanding.
   Q.  Yes, because it was raised by SPMs?
   A.  Correct.
   Q.  And Post Office had known about it since 2012?
   A.  I can't comment on that.
   Q.  Let's move on to the Dalmellington bug please.  This was
       also known as the Branch Outreach issue and again here
       you are heavily reliant on Mr Jenkins who you refer to
       at paragraphs 55, 57, 58 and 61 of your witness
       statement, is that fair?
   A.  Yes.
   Q.  Do you have any first-hand knowledge of this at all?
   A.  Only the research that I have done through this.
   Q.  Let's look at {F/1415} please.  Do you know who prepared
       this document?
   A.  I don't know who prepared it, no.
   Q.  Was it Mr Jenkins?
   A.  I don't know.
   Q.  Look at page 7, please, {F/1415/7}.  So the history of
       the Dalmellington bug -- we can see there are 65
       incidents of this bug in 2010 and 2011 and they are
       broken down there by time, and at the bottom of that you
       will see:
           "Post Office P&BA (FSC) fully aware at the time of
       both Incidents."
           Do you see that?
   A.  I do.
   Q.  And there are two fixes.  There's a 2010 fix, and
       there's a 2011 fix, yes?
   A.  Yes.
   Q.  And if we move forward please now to page 8 {F/1415/8}
       we then have a period from 2011, six incidents; 2012,
       nine incidents; 2013, seven incidents.  Go over the page
       please {F/1415/9}.  2014, nine incidents; 2015, 16
       incidents, and then you will see at the bottom there:
           "Jan 2016 Fix to be applied (subject to POL
       approval)."
           And under all of those years, apart from 2015, the
       last bullet point shows zero calls raised with Fujitsu.
       Shall we go back a slide ?  {F/1415/8}  So zero calls
       raised with Fujitsu in each of those years.  So the
       point that someone was making at Fujitsu there was that
       Post Office had not raised the matter with Fujitsu so
       that it could be corrected earlier.  That's the point
       being made in that slide, isn't it?
   A.  I can't comment.
   Q.  What do you think it looks like?
   A.  It looks like that.
   Q.  So there was no root cause analysis by Fujitsu carried
       out because they had not been contacted by Post Office
       over a period of a number of years, where at least
       identified incidents were numerous.
   A.  I think these incidents came from us going through the
       audit trail to identify where there had been duplicate
       pouches.
   Q.  So SPMs were suffering duplicate pouches over a number
       of years due to a known problem with the system and
       Fujitsu was not informed over that period of years until
       2015?
   A.  As far as I'm aware, the issue was not raised with
       Fujitsu until 2013.  It is entirely possible that
       subpostmasters were dealing with this themselves.
   Q.  Do you mean not raised with Fujitsu until 2015?
   A.  Sorry, I would have to refer back to --
   Q.  If you go over the slide to the next page {F/1415/9} you
       will see the one call is in 2015.  Do you mean 2015?
   A.  I couldn't be categorical on when a call was raised.
       I haven't seen the PEAK or whatever it was that raised
       the call with us.
   Q.  It does not show that any supposed countermeasures to
       ensure the robustness of Horizon as a system were
       working, does it?
   A.  Unfortunately this particular error is not subject to
       receipts and payments problems because it could be -- it
       could be user doing something twice, it's -- the bug had
       the effect of making it look as though a user was simply
       doing something multiple times.
   Q.  Finally, Mr Godeseth, you will be pleased to hear, the
       experts are agreed that Horizon has improved a lot over
       the years.  Would you agree with that assessment?
   A.  I would say that any system improves with time.
   Q.  Would you agree with the assessment of that in relation
       to Horizon?
   A.  Yes.
   Q.  Could you look please at {F/1848.3} to see how incident
       management is now being done, and you can see in the
       "[Incident] detail reported/overview":
           "NBSC called the ATOS service desk and passed on
       a call from Woodgate branch who reported an issue with
       mismatch data between recept and payment in the Trading
       Statement."
           Do you see that?
   A.  I do.
   Q.  And if we go over the page please {F/1848.3/2} under
       "Background", second line:
           "The background to this change was a requirement to
       remove the 'Top Up' options associated to companies that
       should now be branded EE, removing as example the Orange
       & T-Mobile options from all the self-service machines
       and Horizon counters ..."
           Then we go down to the issue:
           "There was a human mistake made when undertaking the
       change to the reference data for these products.
           "The 'front-end' reference was correctly removed.
           "The 'back-end' reference was incorrectly removed at
       the same time."
           At the bottom, final line:
           "This would have shown as a mismatch that the branch
       had to accept."
           That shows us, Mr Godeseth, doesn't it, how branch
       data is incorporated into transaction data and accounts
       for which a mismatch for which an SPM may be held liable
       can arise?
   A.  If the reference data is wrong you can get a receipts
       and payments mismatch.
   Q.  You mean "Yes"?
   A.  You can get a receipts and payments mismatch which would
       be picked up by our monitoring, yes.
   Q.  If we go forward please to an example from 27 November
       it is {F/1842} and if we go forward to page 2 please
       {F/1842/2}, at the top:
           "Enhanced User Management (known as Smart ID in
       branch network) has been introduced to ensure all
       Horizon users are both vetted and have the necessary
       training to transact regulated products."
           Then if we look at the issue, second paragraph:
           "The change was to allow a given EUM user to lock a
       counter and then log on to another counter ..."
   MR JUSTICE FRASER:  I'm sorry, I have lost where you are.
   MR GREEN:  I'm so sorry, I apologise.  On to page 3
       {F/1842/3}, the second paragraph on page 3:
           "The change was to allow a given EUM user to lock
       a counter and then log on to another counter."
           Do you see that?  And then the middle paragraph on
       that page -- in that box:
           "This has caused incidents to occur in the live
       estate ... if the active user session rolls over the
       current stock unit, and a locked session on another
       counter attached to the same stock unit is then resumed.
       This is because when the locked user session is resumed
       the counter is not aware of the rollover and continues
       to trade in the old trading period (TP)/balance period
       (BP).  These transactions are then recorded in the old
       TP/BP and so will not appear on the branch accounts,
       although they will be successfully sent to the back-end
       systems and will be visible in the counter transaction
       log for the old TP/BP:
           "... as of 22 January 19 this has impacted 19
       branches to date over a 6-month period."
           That illustrates a different sort of problem that
       can arise, doesn't it?
   A.  It is a new problem, correct.
   Q.  And it has been impacting branches over a period of
       six months we see there?
   A.  EUM went in fairly recently, so yes, I've got no reason
       to disbelieve that.
   Q.  And finally {F/1848.2}, one from 1 February 2019.  If we
       go to page 2 please {F/1848.2/2}:
           "As part of the Back Office programme to migrate
       from POLSAP to Core Finance System (CFS) a data cleanse
       was completed on any obsolete item IDs."
           Then it mentions Official Postage which would be
       ended and then second paragraph in that background box:
           "Following the item being reinstated overnight on
       the 1st February and being live again on the
       2nd February - NBSC received further calls and it became
       apparent that the accounting sense being applied at the
       counter was positive value not negative value.  We spoke
       to a couple of contacts at Fujitsu but they couldn't
       explain why this was so, we also spoke to Matthew Warren
       who advised that we should try reinstating the default
       mode".
           "What's the issue?":
           "When a branch was processing a claim for Official
       Postage this was increasing the cash on hand figure
       rather than decreasing which could potentially result in
       the branch having a discrepancy if not identified."
           And if we go over the page please to "Impact"
       {F/1848.2/3}, we can see that:
           "As of [2 February] 720 branches are impacted
       including the live kit in NBSC ..."
           Do those sort of problems fairly reflect the
       difficulties that people have to deal with at Fujitsu in
       the normal course of their job?
   A.  Whilst I was certainly aware of the EUM issue, this
       particular one I've got no real knowledge of, or no
       knowledge of.
   MR GREEN:  My Lord, I have no further questions.
   MR JUSTICE FRASER:  Thank you very much.
       Mr De Garr Robinson.
              Re-examination by MR DE GARR ROBINSON
   MR DE GARR ROBINSON:  Mr Godeseth, I would like you to go to
       a document that Mr Green took you to yesterday at
       {F/614}, please.  This is a document by Mark Burley.
       You will see that it is dated 8 April 2010 and you will
       recall that it was the document which charted the
       suspension of the pilot roll-out programme for
       Horizon Online.  Do you remember looking at this
       document?
   A.  I do.
   Q.  And if we could go to page 4 {F/614/4} you will see that
       there were 614 branches live on Horizon, and then on the
       third bullet point it says the high volume pilot was
       suspended.  And if we go on to page 5 {F/614/5}, we will
       see:
           "All migrations suspended as from 25/03."
           There is a reference to the programme support being
       on red alert, and then there's a reference you will see
       in the third bullet point -- four bullet points down:
           "Success Criteria for further migrations under
       discussion with Fujitsu to include:
           "Period of stable running ... FS Review complete and
       key recommendations implemented ... Fujitsu demonstrate
       monitoring and alerting."
           And:
           "All high priority issues resolved or agreed
       workarounds in place."
           From your recollection, can you remember what the
       priority issues were, the high priority issues were?
   A.  For me, my focus was very much on the Oracle bug, so
       I don't have any personal recollection of the other high
       priority issues.  I was establishing myself in Fujitsu,
       I was concentrating on the bug.
   Q.  And what was the Oracle issue?
   A.  I can't remember the details of it, I'm afraid, but it
       certainly involved nodes of the branch database going
       down which had the result that users would see the
       message they were unable to contact the database, or,
       sorry, the data centre.
   Q.  That's fine.  After you were asked questions about this
       you were then shown a PEAK and the document reference is
       {F/588}, perhaps we could look at that.  This is dated
       5 March 2010.  You were taken to this yesterday.  I'm
       not going to ask you any detailed questions on it, you
       will be happy to know, but this was -- as I say, this
       identifies an error experienced during the pilot scheme
       and you were asked a number of questions about it and
       you didn't know much because this wasn't what you were
       working in it at the time.
           Could you explain what the purpose of the pilot
       scheme is?  Why don't you just roll-out Horizon into all
       of the branches all at once?
   A.  That would be frankly a ridiculous approach because you
       had to go through a process of trying out new software
       on a limited set of branches so that you can learn from
       the experience.
   Q.  And when you say learn from the experience, what kind of
       things do you anticipate might happen during an early
       roll-out, even a high volume roll-out of this sort?
   A.  Well, since I recognise this as the occasion that we
       used the branch trading -- sorry, the balancing
       transaction, and I have also noticed from my statement
       that the balancing transaction was to do with £4,000
       being doubled up to 8,000, I think the purpose of
       this -- the purpose of the pilot is simply to go through
       the teething problems that you are bound to have when
       you're introducing a new change.
   Q.  And the PEAK that we see here and the £4,000 problem
       that resulted in the balancing transaction that you have
       just referred to, are those the sorts of problems that
       you anticipate you might encounter during a pilot scheme
       of that sort, or are they not?
   A.  I always expect to encounter problems in the early days
       of any system.  I expect to encounter problems
       immediately after I have put in a change.
   Q.  Thank you, Mr Godeseth.  Now I would like to ask you
       about another document, {F/611}.  This is a PEAK you
       were taken to today, 15 April 2010.  This is also during
       the pilot scheme.  This is in fact the PEAK that led to
       the balancing transaction.  I think it is anyway.
           If we go to page 3 {F/611/3}, you will see --
       perhaps it isn't the balancing transaction one.  You
       will see that -- halfway down the page, 14 April 2010 at
       13:00 hours, Gareth Jenkins says three lines down:
           "What we need to do is the following: ...
           "1.  Update BRDB_BRANCH_STOCK_UNITS where fad_hash =
       ??? AND Branch_accounting_code = 314642 AND stock_Unit
       equals =  ?DEF?  Setting trading_period to 11."
           And 2:
           "Delete BRDB_SU_OPENING_FIGURES WHERE fad_hash = ???
       AND Branch_accounting code = 314642 AND stock_unit =
       ?DEF? trading_period = 12."
           Can I just ask you to explain what the update
       involved?
   A.  The update would simply change a value in a record from
       the 12 to 11, I think in this case, so it would be
       picking a record in the branch database which currently
       has a value of 12 and changing it to 11.
   Q.  And that doesn't -- does that have an impact on the
       branch accounts themselves, on any branch's
       particular -- the values in any branch's accounts?
   A.  It does -- it most certainly does not impact on the
       underlying data.  It is used to put values into the
       correct bucket.
   Q.  And what do you mean by "values into the correct
       bucket"?
   A.  A number of transactions -- so the transactions would
       not disappear, that's the fundamental point.  The raw
       data will not disappear.  It is whether the three
       transactions will appear in TP1, or whether they will
       appear in TP2.  It is a case of which bucket, which --
       ultimately which trading period these transactions would
       be accounted for in.
   Q.  I see.  And the "Delete" instruction, what's the effect
       of the "Delete" instruction?
   A.  It will remove a record from the branch database.
   Q.  And what record was it removing for these purposes?
   A.  This was removing a record which was an opening balance
       for period 12 which happened to have zero in it and --
   Q.  So with removing this record, what happens to the
       balance?  What would the balance have been once this
       zero record was removed?
   A.  Well, this was the initial -- this is the initial record
       needed for period 12, so the problem that was being
       referred to here was that at some point, this record had
       been written into the branch database but then the
       rollover had not completed, so we were left with
       a record in the branch database which shouldn't have
       been there -- sorry, it would have been there had the
       rollover continued, then it and other records would have
       reflected the true position.  What happened here was
       that for some reason, this record was there -- my
       reading of the PEAK says that it was there as a result
       of problems or migration, so where branches were being
       prepared to migrate from old Horizon to new Horizon,
       I think this record was erroneously there.
           So effectively when you then came to do a proper
       rollover, there was a record there and that confused the
       software -- the software had to stop because it found
       a record where there shouldn't be one.
   Q.  And so by deleting that record, what then happens when
       you rollover?
   A.  Once -- then because the record wasn't there, the
       rollover can just happen.
   Q.  Oh, I -- so would I be right in thinking what you're
       saying is that the rollover then happens in the normal
       way based upon the data from the previous transaction
       period as should have happened in the first place; is
       that what you're saying?
   A.  Yes.
   Q.  I see.  Then you were shown the OCP that related to this
       fix.  It's at {F/616.1}.  You have seen this before, OCP
       26361, "Branch 314642 unable to rollover", and then
       about a third of the way down there's a line which reads
       "Steps to implement the fix", do you see that?  And the
       first step is:
           "Logon to BRDB Node 1 as UNIX user 'brdb'."
           Could you explain what "UNIX user brdb" means?
   A.  I don't know any detail, but fundamentally, you're
       logging onto the branch database.
   Q.  And "UNIX user", who would the UNIX user be?
   A.  Oh, sorry, UNIX user would be one of the guys in
       Ireland.
   Q.  One of the guys in Ireland.  So when you say "one of the
       guys in Ireland", do you mean members of the SSC, or do
       you mean someone else?
   A.  Someone else.
   Q.  Well, if we go over the page to page 2 of the same
       document {F/616.1/2} about a third of the way down
       there's a reference to Ed Ashford at Core Services Unix
       Support.  Who is Ed Ashford?
   A.  Ed Ashford is one of the guys in Ireland.
   Q.  Who are the "guys in Ireland" exactly?  Could you just
       clarify -- I appreciate that this may not be your daily
       fare but~...
   A.  They are the people who support the hardware, so UNIX is
       an operating system so they work at a pretty low level
       on the systems.
   Q.  When you say "low level", I mean what do you mean by
       that?  Do you mean they have powerful user rights, or
       they have weak user rights or ..?
   A.  They have pretty powerful user rights, but they are
       very -- very much driven by process as to how they use
       them.
   Q.  What do you mean by "they are driven by process"?
   A.  There's complete and utter control on the processes they
       go through before they do any of these -- this type of
       activity.
   Q.  Okay, well, if I could ask you to go to your witness
       statement, your first witness statement, {E2/1/17}
       please.  I would like you to look at paragraph 59.1.
       You say:
           "A limited number of authorised Fujitsu personnel
       (I am advised by Jason Muir [at] the Security Operations
       team that there are currently 19 at the operating system
       layer and 26 at the database layer) who have access
       privileges that could be used to edit or delete
       transaction data in the BRDB in Horizon Online ..."
           Stopping there, were the guys from Ireland amongst
       these people that you refer to there?
   A.  Yes.
   Q.  I see.  So they are privileged users, are they?
   A.  In my book, yes.
   Q.  And then you say:
           "From the outset, I should make clear that this is
       only a theoretical possibility (which would require the
       circumstance invention of the controls described
       above) ..."
           And then you say:
           "(a) Fujitsu has no policy, process, procedure or
       operational practice that calls for it to use its
       privileged access to edit or delete transaction data."
           And at (b) you say:
           "As far as I am aware, Fujitsu has never used its
       privileged access to edit or delete transaction data."
   A.  Correct.
   Q.  Might it -- given what you have now seen with this OCP
       and the PEAK we looked at before, would you care to
       consider paragraph (b) of your witness statement?
   A.  I think the data that was deleted was not transaction
       data.
   MR DE GARR ROBINSON:  Oh, I see.  My Lord, I have no further
       questions.
                  Questions by MR JUSTICE FRASER
   MR JUSTICE FRASER:  What sort of data would you describe it
       as?
   A.  As I was -- it is data -- sorry, it is a field where you
       are building up a running total, so I would see it as
       analogous to -- if you are adding up 300 different
       numbers, by the time you get to the end of 100 you might
       jot down a total, then you add up the next 100, you jot
       down another total.  So it is where you are building up
       data and this one happens to be an opening balance, but
       it's building up data which is going to be used in
       constructing the end branch trading statement.
   MR JUSTICE FRASER:  So if I wanted to use a phrase for it
       that wasn't "transaction data", what phrase should
       I use?
   A.  I think we have been referring to it as "operational
       data" but "operational" has such a wide connotation
       that ... there possibly ought to be a different
       definition for it.
   MR JUSTICE FRASER:  A different what, sorry?
   A.  There possibly ought to be a different definition for it
       because equally data in the database which is telling
       you which stock unit, you know, what -- which TP you're
       in, whether a stock unit is in use, all of those are
       operational data which do not go down and impact the
       transaction data, so for me, the definition of
       transaction data -- I should be able to get from one
       opening -- one set of opening figures to the end opening
       figures by going back to the transaction data in its
       absolutely raw form.
           As I said before, that is -- fundamentally that's
       not a sensible way to process it, so you do go through
       a sequence of adding things together and then you bring
       them together to actually get the -- to get to your end
       position in a trading statement.
   MR JUSTICE FRASER:  Right.  I have just got a couple of
       questions.  I think you have explained at {E2/1/1} -- we
       will just go to it, it's the very beginning of your
       first statement, your early career.  I think when you
       were instructing in the Navy for those four years, what
       subjects were you instructing?
   A.  Radar and electronics.  Sorry, my Lord, in practical
       terms, that was for the first two and a half years.
       After that I went off to Bureau West in Devizes which is
       where I joined the IT industry.
   MR JUSTICE FRASER:  The IT industry.  Was that while you
       were still in the Navy?
   A.  Correct.
   MR JUSTICE FRASER:  Then you explained about the Impact
       system.  What was your involvement in that, or Impact
       project?
   A.  I was working with Post Office at the time and I was
       interfacing to Fujitsu.  I was also interfacing to Xansa
       who were writing other components involved in the Impact
       programme.
   MR JUSTICE FRASER:  So you were working on the Impact
       programme, were you?
   A.  Correct.
   MR JUSTICE FRASER:  Then if we could call up, please,
       {F/322.2/1}.  I'm so sorry, I have given you a wrong
       reference.  {F/432.2}.  You will remember we looked at
       that this morning.
   A.  Yes, my Lord.
   MR JUSTICE FRASER:  It is an OCP.  If you look against
       "Extra detail" -- Mr Green took you through this in some
       detail already -- but do you see where the new message
       is described?
   A.  I do, my Lord.
   MR JUSTICE FRASER:  There's quantity -- the sale value, that
       would be in pounds, wouldn't it, or effectively
       equivalent to the pounds?
   A.  Yes.
   MR JUSTICE FRASER:  And then "PQty", am I right in reading
       that as "product quantity"?
   A.  I don't recognise the P in there, but I think that would
       be a fair assumption.
   MR JUSTICE FRASER:  All right.  If every single entry is
       a negative, it will depend, won't it, how many entries
       there are as to whether that has a negative or
       a positive effect because if you multiply two negatives
       together you get a positive, or am I being too
       elementary?
   A.  I think this was simply saying that there was an
       original message which had the positive values and then
       you put in another message which has the negative values
       and they will be added together as opposed to
       multiplied.
   MR JUSTICE FRASER:  Well, why would -- if you added quantity
       to product quantity, you would get --
   A.  No, sorry, my Lord, you have -- I think you should see
       these as two records, one above the other, and you're
       going to add the two records together, so the quantity 1
       would be cancelled out by the quantity minus 1, the sale
       value 484 would be cancelled out by the sale value minus
       484 and so on.
   MR JUSTICE FRASER:  So is that describing what effect the
       new message should have rather than what the contents of
       the new message would be?
   A.  I think I would see it more as: this is the record that
       I need to put in --
   MR JUSTICE FRASER:  I see.
   A.  -- to negate the record that's there currently.
   MR JUSTICE FRASER:  So this is the output required
       effectively?
   A.  The output required is all zeros.
   MR JUSTICE FRASER:  Yes.  This is the output required from
       the message to result in all zeros?
   A.  Yes, so my reading of this is there is a record in there
       already, I need to get rid of it logically, so therefore
       I will insert a record which has got all the negatives
       and that will make it go away.
   MR JUSTICE FRASER:  Understood.  Thank you very much.
           Then the next point is at {F/425} we will start at
       page 1 just so you can see what the document is.  Do you
       remember we looked at this document yesterday when we
       were talking about --
   A.  We did, my Lord, yes.
   MR JUSTICE FRASER:  And I think I asked you an Oracle
       question about "Insert" at the beginning of a line.
       Do you remember?
   A.  I do.
   MR JUSTICE FRASER:  If we could look at page 11, please
       {F/425/11}.  Now, the second paragraph there is what
       I asked you a question about, which was where in the
       third line it said "It must start with the 'INSERT INTO'
       clause", do you see that?
   A.  Yes, my Lord.
   MR JUSTICE FRASER:  And I think that's the point that
       I asked you about yesterday, wasn't it?
   A.  Yes.
   MR JUSTICE FRASER:  If we then go down to page 13
       {F/425/13}, and maybe increase the size.  This I think
       is a template for the transaction -- or for the SQL
       statement in the transaction file.  Does that --
   A.  Yes, my Lord.
   MR JUSTICE FRASER:  -- look as if it's right?
   A.  Yes.
   MR JUSTICE FRASER:  And what this page and the one I'm about
       to ask you on the next page is the template for the SQL
       which goes in the transaction file, is that correct?  It
       seems to be from the top two lines.
   A.  So we're inserting into a table called "RX cut off
       summaries."
   MR JUSTICE FRASER:  Yes, but if you look at the very top of
       the quotation which is just after the three lines of
       text it says:
           "The following SQL is a template for the SQL to be
       placed in a transaction file for the~... Transaction
       Correction tool ..."
           Do you see that?
   A.  Yes.
   MR JUSTICE FRASER:  And effectively the template starts with
       the words "INSERT INTO"; is that right?
   A.  Yes.
   MR JUSTICE FRASER:  And if you -- I mean, just in passing,
       the oblique stroke and the asterisk at the beginning of
       the text and the end of the text just before "INSERT
       INTO", those are symbols which mean that the passages
       between the two symbols don't form part of the command,
       is that correct?
   A.  I couldn't be categorical on that, my Lord.
   MR JUSTICE FRASER:  You don't know if that's right -- would
       that be conventional at least?
   A.  That would seem a very sensible convention, yes.
   MR JUSTICE FRASER:  All right.  If you then go on to the
       next page underneath, please {F/425/14} and increase the
       size in that.  That seems to me to be a continuation of
       the sequence which starts on the previous page, is that
       correct?
   A.  Yes.
   MR JUSTICE FRASER:  And where the numerous lines appear
       about eight times one after the other "DO NOT CHANGE
       THIS VALUE FROM THE ONE SPECIFIED!!!", am I right in
       reading those as notes not forming part of the template
       because they don't seem to be in the correct language?
   A.  This to me is an instruction to a user who knows what
       they're doing --
   MR JUSTICE FRASER:  Correct.
   A.  -- to say that the FAD hash has to be -- so you're not
       allowed to go in there and change that.
   MR JUSTICE FRASER:  Exactly.
   A.  You're not allowed to change the branch accounting code.
   MR JUSTICE FRASER:  Yes.  So they don't form part of the
       template, they're effectively warning -- well, I suppose
       they do form part of the template, but they are warnings
       to the person putting the information in the template,
       aren't they?  They're not part of the commands to the
       system?
   A.  Those there are very much warnings to a person who is
       going to use it, who is expected to know what he is
       doing --
   MR JUSTICE FRASER:  Yes, or what she is doing.
   A.  Indeed, my Lord.  And then this is to make it easy, or
       as easy as possible for those people to do the job that
       they are intended to do without making a mistake.
   MR JUSTICE FRASER:  Yes.  Right, and then if we could go on
       to the next page {F/425/15} and increase it, do you see
       there the reference to "bind variables" and the three
       bullet points?
   A.  I do.
   MR JUSTICE FRASER:  Am I right that the bind variables are
       also required to be inputted by whoever is it is putting
       it together?
   A.  I'm afraid I don't know, my Lord.
   MR JUSTICE FRASER:  Do you know who puts the bind variables
       in?
   A.  I don't know who would do that in practical terms, no.
   MR JUSTICE FRASER:  Am I right that somebody would have to
       do that though?  That's not something that seems to me
       to be generated automatically.
   A.  I cannot see how that could be -- if you read the rest
       of the text:
           "Substituted with the values of the branch code
       argument ..."
           That seems -- that to me is telling me exactly as
       you are surmising.
   MR JUSTICE FRASER:  All right, good.  I just wanted to check
       that.  It seemed to me that was what it was doing but
       I just wanted to check.
           Any questions arising out of that?
   MR DE GARR ROBINSON:  My Lord, no.
   MR JUSTICE FRASER:  Any questions?
   MR GREEN:  My Lord, no.
   MR JUSTICE FRASER:  Mr Godeseth, you are now free to leave.
       That's the end of your evidence, thank you very much
       indeed.
   A.  Thank you, my Lord.
                        Procedural matter
   MR JUSTICE FRASER:  Upon reflection, and the fact it is
       after 3 o'clock anyway, I intend not to ask you to call
       Mr Parker.  What I'm going to do is I'm going to give
       some directions in respect of your application --
   MR DE GARR ROBINSON:  My Lord, yes.
   MR JUSTICE FRASER:  -- which has obviously got to be dealt
       with.
           One thing that the email letter that accompanied it
       said which I literally received at 1.55 was that a hard
       copy of it would be brought to court.  Is there a hard
       copy for me?
   MR DE GARR ROBINSON:  My Lord, there is.
   MR JUSTICE FRASER:  Thank you.
           (Pause).
           I think there might be some exhibits.  Is there an
       exhibit as well?
   MR DE GARR ROBINSON:  My Lord, I'm sorry.  (Handed).
   MR JUSTICE FRASER:  Thank you very much.
           Right, so it seems to me that there are three things
       that need to be dealt with in order.  One is any
       response or responsive evidence by the claimants, if
       they oppose the application, which I imagine they will
       need some time to decide, the second is service of
       skeletons, and the third is a date to hear it.
           If it is at all feasible, I would like to hear this
       application on the next day of this trial which is going
       to be Monday 1 April which -- that's only likely to be
       possible if those other two steps, but in particular,
       the first of those two steps, can be accomplished in the
       time between now and then, so that's really a question
       for Mr Green.
   MR DE GARR ROBINSON:  My Lord, it is.
   MR JUSTICE FRASER:  So Mr Green?
                     Submissions by MR GREEN
   MR GREEN:  My Lord, your Lordship won't have seen this
       necessarily but there is a difficulty requiring an
       earlier step, in our respectful submission.
   MR JUSTICE FRASER:  I haven't.  What is it?
   MR GREEN:  If your Lordship has the witness statement of
       Mr Parsons.  My clerks have printed this off over lunch
       so I'm not sure what order your Lordship's one will be
       in, but we have got a witness statement, I think, in
       support.
   MR JUSTICE FRASER:  I have a 14th witness statement of
       Mr Parsons.
   MR GREEN:  Exactly.  If your Lordship would just kindly turn
       to page 6 of that --
   MR JUSTICE FRASER:  I should say, other than reading the
       first couple of paragraphs and then the letter and the
       draft order, I haven't read this witness statement.
   MR GREEN:  That's why I'm drawing your Lordship's attention
       to it.
   MR JUSTICE FRASER:  Right.  Where do we go?
   MR GREEN:  If we go to page 6.
   MR JUSTICE FRASER:  Yes.
   MR GREEN:  The particularisation of the basis of the
       application is I think supposed to be set out in
       paragraphs 24 and 25 essentially.  The first part is to
       say that the findings in relation to matters which arose
       in the common issues trial:
           "... give the clear impression the judge has already
       formed a firm view on these matters.  It is expected
       that this will prevent him from taking an impartial view
       on the same matters when they are revisited at the
       subsequent trials with the benefit of full evidence and
       disclosure."
           And then 25:
           "The judgment also contains a great deal of critical
       invective directed at Post Office, none of which is
       relevant to the determination of the common issues.
       That too creates a clear impression that the judge has
       not behaved impartially.  The same can be said for those
       parts of the judgment which harshly criticised
       Post Office's witnesses on matters irrelevant to the
       common issues."
   MR JUSTICE FRASER:  Yes.
   MR GREEN:  And, my Lord, those complaints are wholly
       unparticularised, they are serious matters and if we are
       going to analyse them and respond to them -- I've never
       seen any -- I have not seen very many recusal
       applications, but I have not seen a recusal application
       as lacking in particularity as this.  If there are
       specific criticisms of matters in the judgment that are
       said to show -- create an impression of your Lordship
       not having behaved impartially, any applicant in that
       situation must set those out with particularity, for two
       reasons: so that your Lordship can fairly consider what
       is being said, not in some vague penumbra of complaint
       by the losing party, but also so that the claimants have
       an opportunity to consider precisely what the basis of
       this remarkable application actually is, and at the
       moment, we don't actually know which parts of the
       judgment they are pointing to.
           What is said to be -- what is a fair appraisal of
       the credibility of a particular witness and what is
       invective?  And it's not to be left for us to scrabble
       around, as we have had to do on some other things, to
       try and work out what's being said, and so, my Lord, at
       the moment we're not in a position to form a view on the
       application, less still file responsive evidence.
           The allegations -- the basis of this application,
       which is a serious application to make, and is likely if
       not calculated to derail these proceedings, has to be
       spelt out with proper particularity before we are in
       a position to respond to it, so I would invite
       your Lordship to consider directing that that be done
       and I would invite your Lordship to do so within
       a fairly short time period, because no solicitor signing
       a witness statement of this sort could put his name to
       an allegation of that sort without having satisfied
       himself of precisely the basis upon which it was being
       said.  So this is not "Oh, I will go away and work it
       out now"; this must have been worked out with great care
       already, otherwise an allegation of that sort could not
       properly be made.
   MR JUSTICE FRASER:  All right, well, just before I ask you
       for a response on that, Mr De Garr Robinson, what I'm
       actually going to do is I'm going to go and read the
       whole witness statement, because I think I quite clearly
       have explained I haven't had a chance to read it yet.
   MR DE GARR ROBINSON:  Nor have I.
   MR JUSTICE FRASER:  Well, you might then also benefit from
       having had a chance to look at it.
   MR DE GARR ROBINSON:  Indeed.
   MR JUSTICE FRASER:  I will come back in at I think 3.40.
       I will take all the papers with me.  The only ones that
       I will probably have had a real chance to look at in
       great detail is the witness statement.  I know the
       judgment fairly well, for obvious reasons.  Then I will
       come back, I will hear Mr De Garr Robinson what you have
       to say about Mr Green's point, and we will revisit the
       whole question of timetabling if I have missed out a
       important step which, upon reflection, I decide ought to
       be taken.  So 3.40.
   (3.14 pm)
                          (Short Break)
   (3.44 pm)
   MR JUSTICE FRASER:  Mr De Garr Robinson.
                Submissions by MR DE GARR ROBINSON
   MR DE GARR ROBINSON:  My Lord, responding to my learned
       friend's submission, he puts his point as a requirement
       for particulars.  The truth is this is a matter of
       argument, and your Lordship and indeed my learned
       friend, have already had no shortage of arguments.
       There was an overwhelming number of those arguments at
       the common issues trial.
           It is not necessary for there to be a list of
       critical findings in your Lordship's judgment, everybody
       knows what those findings are and my learned friend's
       suggestion that he needs particulars in order to decide
       his position on the application is frankly surreal.
           However, my Lord, I do have a proposal which in my
       submission will be better than particulars, which is to
       order sequential skeleton arguments and I would invite
       your Lordship to order that the defendant's skeleton
       argument be served by close of business on Tuesday, that
       the claimants' skeleton argument be served by close of
       business on Thursday, so that everyone is ready for the
       hearing that you suggested, that obviously should be
       dealt with as quickly as possible, on Monday the 1st.
   MR JUSTICE FRASER:  I'm sorry, when did you say for the
       claimants'?  End of Thursday?
   MR DE GARR ROBINSON:  Thursday.
   MR JUSTICE FRASER:  Well, that proposal doesn't include
       within it provision for any evidence that they might
       want to put in?
   MR DE GARR ROBINSON:  My Lord, that's true.
   MR JUSTICE FRASER:  And I think it is also -- it effectively
       takes as read that they are going to oppose the
       application.
   MR DE GARR ROBINSON:  Well, my learned friend suggested
       otherwise, which is why I --
   MR JUSTICE FRASER:  I think he said he -- but in a way ...
       yes, all right.  So you say no need for any evidence in
       response.
   MR DE GARR ROBINSON:  Well, my Lord, if my learned friend
       wishes to put in evidence, he can put in evidence, but
       I do respectfully suggest that when he talks about
       particulars, he is really talking about argument, and
       I have no objection to giving him prior sight of --
       I should say the defendant has no objection to giving
       the claimants prior sight of their arguments, but,
       my Lord, to slow matters down by having the formality,
       in my respectful submission, the useless and expensive
       formality of a list of points -- I don't quite know what
       the particulars would look like -- that wouldn't add to
       the efficiency of this process and it certainly wouldn't
       add to its speed.
   MR JUSTICE FRASER:  No, but it is important that other
       matters aren't sacrificed in the interests of speed,
       given the nature of the application.
   MR DE GARR ROBINSON:  Of course.
   MR JUSTICE FRASER:  If the skeleton -- the defendant's
       skeleton argument were to identify in a schedule the
       following three points -- because apart from anything
       else, I will need to know this at some point: the
       specific findings that are referred to in paragraph 24,
       the critical invective referred to in paragraph 25 and
       the harsh criticisms referred to in paragraph 25, then
       that would effectively -- whether it comes in a schedule
       to a skeleton, or whether it comes in the 15th witness
       statement of Mr Parsons, that would effectively identify
       what parts of the judgment were being attacked -- or
       were being relied on to found the application.
           But let me hear from Mr Green about the sequential
       skeleton approach.
   MR DE GARR ROBINSON:  My Lord, it would, but may I just
       mention one thing.
   MR JUSTICE FRASER:  Yes.
   MR DE GARR ROBINSON:  Were an order -- a direction to be
       made requiring a skeleton which contains a schedule of
       that sort, those things would not be things in relation
       to which my learned friend would wish to put in evidence
       in response.
   MR JUSTICE FRASER:  I entirely accept that, but what they
       would do, for example, Mr De Garr Robinson, just taking
       paragraph 24 as an example, they would say "These are
       the specific findings which we, the Post Office, says
       justifies the suggestion that the judge has already made
       a clear -- that justify the clear impression that the
       judge has formed a firm view on these other matters."
       Now, those passage are going to have to be brought to my
       attention at some point.
   MR DE GARR ROBINSON:  Absolutely.  My Lord, I'm not
       resisting any direction of that sort, just to be clear,
       but what I am resisting is the suggestion that there
       needs to be some further time to allow evidence to be
       formulated once these matters are flushed out.
   MR JUSTICE FRASER:  Evidence arguing to the contrary.
   MR DE GARR ROBINSON:  Yes.
   MR JUSTICE FRASER:  I agree with you about that.  As
       I understand the witness statement of Mr Parsons, which
       I have now read very carefully, twice, it is based on
       the judgment, number 3 on the common issues, or
       certainly that's how it is explained in the actual
       witness statement, and insofar as it is necessary to
       look at other things, they are things in respect of
       which there is a transcript available because they
       relate to matters that have happened at an interlocutory
       hearing, so there's not going to be any evidential
       argument about what actually happened.  Is that
       a correct analysis?
   MR DE GARR ROBINSON:  My Lord, all I have done is read the
       witness statement in the same way that your Lordship
       has.
   MR JUSTICE FRASER:  All right.  Well, I don't want unfairly
       to put you on the spot about that.
           Mr Green.
                     Submissions by MR GREEN
   MR GREEN:  My Lord, I couldn't respectfully disagree more --
   MR JUSTICE FRASER:  With?
   MR GREEN:  -- with what Mr De Garr Robinson has just said.
           The starting position for any application of this
       type is that it must immediately be made on a basis
       which clearly sets out, with particularity, the aspects
       of the judgment in this case and findings which are said
       to demonstrate apparent bias of the court.  That is
       a totally uncontroversial statement and one that is
       not -- a principle that's not lightly to be departed
       from.  And the reason for that is it affords the party
       on the other side an opportunity to consider, with
       clarity, whether they wish to resist the application or
       not.
   MR JUSTICE FRASER:  Yes.
   MR GREEN:  Before they are committed to a 24-hour or 36-hour
       or 48-hour window in which to respond to a skeleton
       argument, a skeleton argument which has doubtless
       already been considered insofar as the matters which
       would go into the schedule, and will continue to be
       considered without us knowing what those points are all
       the way up until Wednesday when it is served and when we
       then learn for the first time what the actual basis of
       this application is.
           My Lord, the reason I say that is this: a finding
       which is adverse to a party is not a proper foundation
       for an application of this sort.
   MR JUSTICE FRASER:  Yes.
   MR GREEN:  A criticism of a witness is not a proper
       foundation for an application of this sort.  And
       therefore, it is essential that the applicant in this
       application spells out, in every case, what is said to
       be critical invective going beyond the bounds of what
       a court may properly do without creating the apparent
       bias now alleged and what it is -- which findings it is
       said or parts of the judgment which harshly criticise
       Post Office witnesses on matters irrelevant to the
       common issues.
           Those matters have been to be spelled out and
       my learned friend, Mr De Garr Robinson, has just told
       the court that everybody knows which parts of the
       judgment he is referring to.  If that's true --
       I don't -- but if there's any basis for saying that to
       the court, then he can write them out in a list and,
       my Lord, this application --
   MR JUSTICE FRASER:  Well, I'm not necessarily sure
       Mr De Garr Robinson could write them out in a list
       but --
   MR GREEN:  Sorry, your Lordship is quite right.
   MR JUSTICE FRASER:  -- somebody at the Post Office can write
       them out.  Mr De Garr Robinson has been dealing with the
       computer issues trial.
   MR GREEN:  My Lord, that's a very fair point and I should
       observe that distinction, but it doesn't take away from
       the fact that whoever has prepared and advised on and
       considered whether this is a proper application to make
       must have identified what the basis is, so asking them
       to spell it out is not a subsequent step, they must have
       done that before this could be signed off.  This is not
       an application for directions or something like that,
       this is a very important application and it cannot have
       been made without carefully thinking what the elements
       in each case are and particularly where it is said that
       the criticism, for example, of Post Office's witnesses
       on matters irrelevant to the common issues -- that has
       to be spelled out too and its basis.  It is not
       satisfactory, in our respectful submission, for this to
       be done the other way round, so that we only find out
       after a skeleton argument has been filed in an
       application what the basis is and we then have to decide
       both whether to resist it and on what basis within
       a very truncated time period.
           So I would be inviting your Lordship to order
       provision of particulars of those two paragraphs, by
       list, and an explanation, a proper explanation of the
       basis of this application.
   MR JUSTICE FRASER:  Well, the second of those -- what I'm
       minded to do -- and I will explain what it is and
       Mr De Garr Robinson can make any submissions about it he
       wants to.  What I am minded to do is order the
       Post Office to provide a witness statement by noon on
       Tuesday 26th -- but we can come and discuss the timing
       in a moment -- that sets out: 1, the specific findings
       in the judgment referred to in paragraph 24 of
       Mr Parson's 14th witness statement; 2, the critical
       invective referred to in paragraph 25 of Parsons 14; and
       3, the criticisms of the Post Office witnesses referred
       to in paragraph 25.
           Now, in a way, Mr De Garr Robinson, it doesn't much
       matter whether that's a schedule -- it would be the same
       exercise whether it was a schedule to a skeleton or
       whether it was in another witness statement from
       Mr Parsons, but I think in the circumstances it should
       be in a witness statement.
                Submissions by MR DE GARR ROBINSON
   MR DE GARR ROBINSON:  Well, my Lord, I would resist that.
       It shouldn't be in a witness statement in my respectful
       submission.  It's a list of particulars and the
       appropriate document would be an order for particulars.
       Actually it would be preferable, and more efficient in
       my submission, if a skeleton argument were directed
       which should have a schedule setting out those
       particulars.  That would actually be beneficial to --
       and more useful for the claimants than a witness
       statement or some independent document that goes on --
   MR JUSTICE FRASER:  Well, I think everyone is agreed there
       has to be an identification of them.
   MR DE GARR ROBINSON:  Yes.
   MR JUSTICE FRASER:  In these circumstances I think it is
       best if it is in a witness statement, but if that is to
       refer to an exhibited schedule, well that's one way of
       doing it.  I'm not going to debate it endlessly.
   MR DE GARR ROBINSON:  Well, my Lord, could I just address
       your Lordship on that.
   MR JUSTICE FRASER:  Yes.
   MR DE GARR ROBINSON:  There is -- it would put the person
       making the witness statement in an invidious position.
       I mean, if your Lordship is happy to direct a witness
       statement in which a human being has to make a witness
       statement saying these things then of course that order
       will be complied with.
   MR JUSTICE FRASER:  Why would it put them in an invidious
       position?
   MR DE GARR ROBINSON:  Well, my Lord, because it will be
       a human being criticising a High Court judge for making
       findings of that sort.  Bearing in mind that the
       important thing is simply that the information is
       provided so that it can be understood, in my respectful
       submission it would be better for it to be in a schedule
       to a skeleton argument.  If your Lordship wishes to
       order that a human being makes a witness statement, then
       that will of course be complied with.
   MR JUSTICE FRASER:  Well, the purpose of this is not to
       personalise it or in any way identify somebody as being
       the personification of the application.  I was assuming
       it would be Mr Parsons because he has said these things
       already in Parsons 14.
   MR DE GARR ROBINSON:  He has.
   MR JUSTICE FRASER:  So it would simply be identification of
       which those passages are -- I don't think it is
       necessary to add this, but I will in any case -- by
       explaining which passages to which he is referring in
       his 15th witness statement, or it might even be the
       first witness statement of someone else.  I don't think
       there's any concern about exposure of a human being
       because it would be coming from one of the parties'
       advisors.
   MR DE GARR ROBINSON:  I have made my submission.  I'm not
       sure I can helpfully add to it.
   MR JUSTICE FRASER:  And if it is -- and even if it is
       a simple form of words like "The findings to which
       I refer in paragraph 24 are ... 1, 2, 3, 4, 5, 6, 7 ..."
       I mean, I think that would be satisfactory but I'm not
       going to give any specific direction about what the
       contents of it should be.  I think it's just explaining
       what the paragraphs or the findings or the parts of the
       judgments are which have already been referred to in
       paragraphs 24 and 25 of Parsons 14.
   MR DE GARR ROBINSON:  I understand.  My Lord, does that mean
       that the hearing of the application will be delayed
       beyond Monday?
   MR JUSTICE FRASER:  Well, that's what we're going to have to
       address now I think because what we have to do, or what
       I have to do, is I have to balance enough time for the
       steps that are necessary to be taken properly without
       simply imposing inordinate delay on the process
       generally, so if noon on Tuesday for the witness
       statement -- by then that would be Tuesday 26th -- then
       the claimants are going to have to provide their
       response -- or identify, firstly, if they oppose the
       application or not by a particular time or date which we
       will come on to.  Then if so advised, I suppose
       technically they have to be given the opportunity to put
       evidence in response, but I can't imagine if the
       substance of the application is the judgment that would
       necessarily be a step that would be taken, but I can't,
       I don't think, press Mr Green to tell me now whether or
       not it definitely will be, and then we will have to put
       some provisions in for skeletons which, given these
       circumstances, can be simultaneous I think.
           So, Mr Green, if you get either Parsons 15 or Mr or
       Mrs X number 1 by noon on Tuesday, when would you like
       firstly to notify the other side of your position in
       respect of the application?
   MR GREEN:  My Lord, I think we would have to have 24 hours.
   MR JUSTICE FRASER:  Yes, all right.  So if you're asking
       until noon on the 27th?
   MR GREEN:  My Lord, yes.
   MR JUSTICE FRASER:  And I imagine that will just be in
       a letter?
   MR GREEN:  Indeed.
   MR JUSTICE FRASER:  Then if so advised, evidence in response
       by ..?
   MR GREEN:  Hopefully we would be able to do that within a --
   MR JUSTICE FRASER:  Friday?
   MR GREEN:  Yes.
   MR JUSTICE FRASER:  By noon on Friday?
   MR GREEN:  By noon on Friday.  My Lord, I think --
   MR JUSTICE FRASER:  Just before I get onto skeletons,
       Mr De Garr Robinson, do those two intervals require any
       observation from you?
   MR DE GARR ROBINSON:  My Lord, none.
   MR JUSTICE FRASER:  All right.  That then means the
       hearing -- I think I said rather peremptorily before
       I left court it would be 1 April, but I think it's going
       to have to be a day or two later in that week because
       you're going to have to have provisions for skeletons,
       I'm then going to have to have a little time to read
       them, and the application says set down for four hours
       but I think in the circumstances, I will just set it
       down for a day.
           Skeletons by 10 o'clock on Tuesday 2nd with the
       hearing on Wednesday 3rd?
   MR DE GARR ROBINSON:  My Lord, I'm unaware whether
       Lord Grabiner will be available on Wednesday and
       enquiries need to be made.
   MR JUSTICE FRASER:  Well, Mr De Garr Robinson, in all the
       circumstances and given that is actually one of the days
       this trial should have been sitting, if you want to move
       it from the 3rd you're going to have to issue an
       application I'm afraid because I'm not going to say
       anything more that at this point.  So skeletons at the
       end of Monday 1st, let's say 5 o'clock Monday 1st, the
       skeletons, together with a single bundle of authorities,
       although I can't imagine any of them will be
       controversial, the test is fairly well-established.
           Any further directions are going to have to wait
       until the application has been resolved, I think.  Will
       the parties draw up an order?
   MR DE GARR ROBINSON:  My Lord, yes.  When you say "any
       further directions", do you mean directions in relation
       to the Horizon trial?
   MR JUSTICE FRASER:  Well, not necessarily in relation to the
       Horizon trial, because that would suggest I have
       prejudged the outcome of the application which obviously
       I haven't.  Well, the question of whether I would
       reserve judgment, how long it would take, all of those
       sorts of things, I don't think I'm in a position to say
       anything about that now until having heard -- seen the
       evidence and heard the arguments on the 3rd.
   MR DE GARR ROBINSON:  Yes.
           (Pause).
           My Lord, we are not presently aware of what
       Lord Grabiner's availability will be, but we should know
       by Monday, and if an application needs to be made,
       your Lordship will hear it then.
   MR JUSTICE FRASER:  Although I'm not making any directions
       about this, it would be foolish to ignore the fact that
       we still have got in the Horizon issues trial definitely
       Mr Parker, who I think is going to be less than half
       a day, because you haven't got longer than that anyway.
   MR GREEN:  No.
   MR JUSTICE FRASER:  The position vis-Ă -vis Mr Membery, which
       I know is in a state of flux because of his personal
       circumstances.
   MR DE GARR ROBINSON:  Yes.
   MR JUSTICE FRASER:  And then the experts.  I think it is
       pretty clear we won't be able to hear any more -- if the
       application is not acceded to, we won't be able to hear
       any expert evidence at all this side of Easter --
   MR DE GARR ROBINSON:  I see.
   MR JUSTICE FRASER:  -- it seems to me, because apart from
       anything else, depending on the outcome of the
       application, either party might want to make a new
       application about further delay, I say, expressing
       myself neutrally, because they might want to appeal
       whatever my answer is, but we will cross those bridges
       if and when we come to them.
           So I think that deals with everything in respect of
       the application to recuse.  Is there anything I have
       missed out?
   MR DE GARR ROBINSON:  My Lord, not that I can think of.
   MR JUSTICE FRASER:  Anything I have missed out?
   MR GREEN:  My Lord, no.
                           Housekeeping
   MR JUSTICE FRASER:  Turning to the rather more mundane
       subject of outstanding housekeeping in what we have
       already been doing, I was still waiting for hard copies
       of the PEAKs and the KELs.
   MR GREEN:  My Lord, yes, we're going to put those together
       for you and deliver them tomorrow, if that's all right.
   MR JUSTICE FRASER:  I think you should probably still do
       that --
   MR GREEN:  It has to be done.
   MR JUSTICE FRASER:  -- so it doesn't get lost -- well, it
       will have to be done for whoever it is who is the trial
       judge who takes the matter all the way through, whether
       it is me or somebody else.
           There was still a hard copy of Dr Worden's report
       with appendices.
   MR DE GARR ROBINSON:  Oh, has that not got to you?  I do
       apologise.
   MR JUSTICE FRASER:  It might have appeared somewhere, but
       I don't believe it has.
           I did, however, get three hard copies of the
       application to recuse myself as the managing judge, but
       I didn't get any hard copy of Dr Worden.
   MR DE GARR ROBINSON:  I'm so sorry.
   MR JUSTICE FRASER:  And then there was still an outstanding
       summary from you, Mr De Garr Robinson, about your
       privilege redaction review but in the circumstances
       I don't need to pursue that with you now.
   MR DE GARR ROBINSON:  That's very kind.  I am prepared to
       address your Lordship on it now should you wish, but I'm
       happy to -- perhaps it would be better ...
   MR JUSTICE FRASER:  I think we will put that on the -- as
       the only matter still outstanding so far as current
       housekeeping is concerned in this trial, until the
       application has been dealt with.
   MR DE GARR ROBINSON:  Very good, my Lord.
   MR JUSTICE FRASER:  Anything else?
   MR GREEN:  My Lord, no.
   MR JUSTICE FRASER:  So someone is going to draw up an order.
   MR DE GARR ROBINSON:  We will decide, my Lord.
   MR JUSTICE FRASER:  I think you can probably agree between
       yourselves who that is, and until Wednesday 3rd at 10.30
       and we will have Opus here for that obviously.
   MR GREEN:  My Lord, yes.
   (4.07 pm)
             (The court adjourned until 10.30 am on
                     Wednesday, 3 April 2019)