Wednesday 13 March 2019

Horizon Trial: Day 3 transcript

This is the unperfected transcript of the third day of the Horizon trial, held on 13 March 2019 in court 26 of the High Court's Rolls Building.

                                       Wednesday, 13 March 2019
   (10.30 am)
   MR HENDERSON:  My Lord, a couple of housekeeping points
       first of all.  I reviewed my notes for cross-examination
       overnight, particularly in the light of your Lordship's
       final observation of the suggestion that various points
       be put to Mr Patny junior as well and I have taken the
       view that I don't need to continue my cross-examination
       of Mr Patny senior.
   MR JUSTICE FRASER:  Understood.
   MR HENDERSON:  So I have notified my learned friend
       yesterday evening, but apologies for that slight change.
   MR JUSTICE FRASER:  No, it is not an apology.  So you don't
       have any re-examination?
   MR GREEN:  I don't, my Lord.
   MR JUSTICE FRASER:  Let me just check if I have any
       questions for him.  Just give me one second.
           There was one minor point but it's not worth having
       him back to ask him.  So we're going to start with a new
       witness today.
   MR HENDERSON:  I'm grateful.  And just for your Lordship's
       information, I think we're on track in terms of
   MR JUSTICE FRASER:  Don't worry about that.  I have been
       keeping an eye on that.  Right, so Mr Green -- sorry,
       you had some more housekeeping?
   MR JUSTICE FRASER:  Have you got housekeeping?
   MR GREEN:  Only one thing just in case your Lordship wants
       to see it.  Mr Roll wanted to clarify more precisely
       some points in his witness statement overnight and in
       accordance with what we did in the common issues trial
       we have incorporated it into his statement so
       your Lordship can see how it fits in.
   MR JUSTICE FRASER:  If you could hand that up that would be
       excellent.  Thank you.  So he is being called on
   MR GREEN:  He is going to be called this afternoon
       I understand because we're going to -- I understand
       we're going to get to him this afternoon.
   MR JUSTICE FRASER:  All right, fine.
   MR HENDERSON:  Actually there was one final matter I was
       going to deal with during cross-examination but perhaps
       it is more convenient to deal with it now.  There is
       a further document which is on its way on to finding its
       way to the trial bundle but isn't there yet.  So I have
       hard copies again, I have notified my learned friend,
       but could I perhaps hand a copy up just for convenience
   MR JUSTICE FRASER:  Yes please.  If it hasn't been uploaded
       we can use it in the old fashioned way, that's not
       a problem.
   MR HENDERSON:  Yes of course and I will hand it to the
       witness in due course.
   MR JUSTICE FRASER:  All right, I will just put that to one
           On the basis that we're dealing with housekeeping
       generally, I was going to deal with this point at the
       end of today but I will deal with it now and a similar
       point came up in the common issues trial, or two similar
       points came up, but I will just explain what they are.
       If you're going to be putting -- and you are of course
       perfectly entitled to put certain matters which might
       suggest criminal conduct, at an appropriate point the
       witness has to have the warning against
       self-incrimination.  If you can just bear that in
       mind -- it doesn't have to be done at the very beginning
       of their evidence, it can be done at the stage when you
       get to that cross-examination, but can you just bear
       that in mind so that the warning has to be administered.
   MR HENDERSON:  My Lord, yes.
   MR JUSTICE FRASER:  You don't have to tell me in advance,
       but each witness is potentially, if that's going to be
       put to them they have to be given that warning.  That's
       the first point.
           The second point is as far as what we have called in
       the common issues trial a hard start in terms of trial
       timetabling, if you find yourself going quicker than you
       expected, that is not a problem.  I am working on the
       basis that we will be starting with the Post Office
       witnesses on Monday morning in terms of a hard start for
       week 2.  I am not in any way going to become intemperate
       or get upset if for example on Thursday at 11.30 or
       early afternoon you effectively run out of questions to
       ask the Post Office witnesses and the same applies for
       the witness of fact.  We're basically doing this trial
       in self-contained weeks.
           Then the final point, which is really a point for
       Mr Draper: I reviewed the transcript overnight, it
       appeared to me that I came across as an exceptionally
       abrupt in at least two occasions in respect of the way
       you were cross-examining Mr Latif, that was really in
       the sense that (a) I was aware of the time difference
       and (b) I was aware of the fact that he was on
       a videolink from Pakistan, which is why I was probably
       a little more -- or I might have come across a little
       more acerbic about how long you were going to be and
       I don't want you to read anything into that.
   MR DRAPER:  I understand.
   MR JUSTICE FRASER:  All right?  So is that all the
   MR HENDERSON:  My Lord, yes.  Dealing with the first point
       first which is the relevant warning, Mr Aakash Patny is
       coming in this morning and I think out of an abundance
       of caution it may simply be sensible to make the warning
       since we are talking about a lot of declarations.
   MR JUSTICE FRASER:  Right, okay.  Do you have it with you,
       Mr Green?
   MR GREEN:  My Lord, I haven't brought it with me and
       I should have done this morning because your Lordship
       didn't have the wording yesterday.  I apologise.
   MR JUSTICE FRASER:  It will be in the White Book.
   MR GREEN:  We can find it on the transcript of the common
       issues trial which we've got on Opus.
   MR JUSTICE FRASER:  Or it will be in volume 2.  So you do
       Opus, I will do volume 2 and we will see who gets there
   MR GREEN:  My Lord, I'm afraid you are going to beat me
       because I can't get back in from this Opus to what
       I have in chambers.
   MR JUSTICE FRASER:  I wouldn't give up so quickly, you don't
       know how long it is going to take me to do it this way.
       Just remind me of the section?
   MR GREEN:  Is it C8.11 in -- have we got it?  My learned
       junior Mr Miletic has found it in the transcript.
   MR JUSTICE FRASER:  Oh, he has found it in the transcript.
       All right, then just call it up because I've got to ...
   MR GREEN:  So in the documents bundle --
   MR JUSTICE FRASER:  I have found it, it is 9B, 10.42.
       "Privilege against incrimination", section 14.  So if
       you call up what I said last I think.
   MR GREEN:  My Lord, I've got it at the wording that is --
       shall I just read it out to your Lordship?
   MR JUSTICE FRASER:  Where are you looking?
   MR GREEN:  So this is the warning your Lordship gave
       Mrs Stockdale.
   MR JUSTICE FRASER:  Why can't it just be called up on the
   MR GREEN:  It can be called up on the screen.  So it is
       {C8.11/24/50} and it is on internal page 196 and it
       begins at line 6.
   MR JUSTICE FRASER:  Or line 7.
   MR GREEN:  And it is line 10 where your Lordship explains
   MR JUSTICE FRASER:  Yes.  Now, last time I said to the
       witness what the specific offence was.  I'm not sure
       that necessarily is required, so instead of specifying
       that false accounting was a criminal offence I am going
       to be saying what?  "Certain facts and matters in
       relation to your making of cash declarations"?
   MR HENDERSON:  Yes, "making declarations of cash and stock".
   MR JUSTICE FRASER:  "Making declarations to the Post Office
       regarding the cash and stock."
   MR HENDERSON:  My Lord, yes.
   MR JUSTICE FRASER:  Right, well, I will do that at the
           Is that everything?
   MR JUSTICE FRASER:  Yes?  Are you about to call another
   MR GREEN:  I am, my Lord.
   MR GREEN:  My Lord, I am calling Mr Aakash Patny.
                    MR AAKASH PATNY (affirmed)
   MR JUSTICE FRASER:  Thank you very much, Mr Patny.  Have
       a seat.  Can I firstly say that you have been referred
       to once or twice as Mr Patny junior.  That's not meant
       to be patronising or anything like that, it was just to
       differentiate you from your father.
           Secondly, I'm now going to give you a warning which
       you have heard debated.  I think you might have been in
       court yesterday?
   A.  Yes.
   MR JUSTICE FRASER:  So you would have heard a similar form
       of words to your father.  But it is just to tell you,
       Mr Henderson is going to ask you some questions and some
       of those questions are going to relate to certain facts
       and matters relating to the declarations that you made
       to the Post Office regarding cash and stock.  Under the
       Civil Evidence Act 1968 you have the right to refuse to
       answer any of his questions if it would tend to expose
       you to proceedings for a criminal offence or the
       recovery of a penalty.  So you might choose to answer
       his questions but it is correct that I should warn you
       you don't have to answer any of them if you don't want
       to.  If he asks you a question and you want to decline
       to answer it, just say "I decline to answer that
       question".  Is that fairly clear?
   A.  Yes.
   MR JUSTICE FRASER:  Thank you very much, Mr Patny.  Over to
       you, Mr Green.
                 Examination-in-chief by MR GREEN
   MR GREEN:  Mr Patny, hopefully there is a bundle in front of
       you and if you turn to tab 2 please {E1/2/1}.  You will
       see there "Amended witness statement of Aakash Patny"?
   A.  It has Adrees Latif on it.  Sorry, wrong tab sorry.
   Q.  That's it, if you're in tab 2 do you have "Amended
       witness statement of Aakash Patny"?
   A.  Yes.
   Q.  And if you turn very kindly to the back of that
       statement, page 4 {E1/2/4}.
   A.  Yes.
   Q.  Do you have that?  And is that your signature?
   A.  It is, yes.
   Q.  And do you believe the contents of the statement to be
   A.  Yes.
   Q.  Thank you very much.  Would you wait there.  There will
       be some questions.
                Cross-examination by MR HENDERSON
   MR HENDERSON:  Mr Patny, good morning.
   A.  Good morning.
   Q.  I assume that you were in court yesterday when your
       father gave evidence, is that right?
   A.  Yes.
   Q.  And I don't want to go through everything with you that
       I said to him, but there are a few points I would like
       to deal with.
   A.  Okay.
   Q.  In particular I want to deal with various allegations
       that you make concerning a shortfall, which varies in
       amount but which during May 2016 comes and goes, in the
       value of about £17,000.  Do you recall that?
   A.  Yes.
   Q.  And eventually it is right, isn't it, that a sum of
       around that sum of money was required by the Post Office
       to be settled centrally.  Do you recall that?
   A.  Yes.
   Q.  Although it hasn't actually been paid to the Post Office
       I think?
   A.  No, it hasn't.
   Q.  And that was towards the end of May or the beginning
       of June, something like that?
   A.  Around that time, yes.
   Q.  And as I understand the case put forward by you and your
       father is that this was caused by a bug or bugs in
       Horizon, is that right?
   A.  I believe so, yes.
   Q.  And your evidence is that the outage on 9 May is the
       start of the problems, is that right?
   A.  Yes.
   Q.  Okay.  And again -- and I think this is just covering
       common ground and you have dealt with it in your
       statement -- I think you are responsible for the
       declarations in the branch, or you were responsible for
       the declarations in the branch?
   A.  Mostly, yes.
   Q.  Cash and stamps particularly?
   A.  Yes.
   Q.  But it is also right, and your father explained
       yesterday that you sometimes shared log in IDs, is that
   A.  Sometimes, yes.
   Q.  So I hope I can take this reasonably quickly.  I don't
       want to go through with you the events of the outage.
       I discussed that with your father yesterday as you will
       recall, but can we look at {F/1507.1} and can you enable
       editing and remove filtering.
   MR JUSTICE FRASER:  What document is this?
   MR HENDERSON:  This is the event data I think for the
       relevant branch and it shows -- well, can we go to
       row -- it is the event data spreadsheet.  Can we go to
       row 13904.  We looked at this yesterday.
           Mr Patny, it would help me -- when your father was
       being cross-examined yesterday were you able to see the
       documents on the screen from somewhere?
   A.  No.
   Q.  Not particularly, okay, fair enough.
           So this is the day of the outage and this is the
       cash declaration that you made at the end of that day at
       19.04.  Do you see that?
   A.  Yes.
   Q.  And below it we can see that there is a variance for
       minus £1,138.21, do you see that?
   A.  Yes.
   Q.  And that suggests that there is less cash than you were
       expecting to the tune of that figure?
   A.  Yes.
   Q.  So based on that, do you agree that it is very unlikely
       that the outage that had occurred during that day was
       responsible for the discrepancy of £17,000?
   A.  No.
   Q.  Why not?
   A.  Well, that's a big amount to be missing, isn't it,
   Q.  Yes, but it's not £17,000, that's my point.
   A.  No, it's not £17,000, no.
   Q.  My understanding of your evidence is that you believe
       that the outage on 9 May led to the discrepancy of
       £17,000 which you discovered on 11 May.
   A.  Yes.
   Q.  And what I'm taking you to is the position on 9 May on
       the day of the outage itself and pointing out that there
       wasn't a discrepancy of £17,000 on that day.  Do you
       accept that?
   A.  I do, but it is after 7 o'clock as well so that's after
       polling time, so I don't know if that could have been an
       issue as well.
   Q.  Why would that have been an issue?
   A.  I'm not too sure.
   Q.  You discovered, or you tell us you discovered the
       shortfall on 11 May.  Can we have a look at {F/1522.1}.
       This is the help desk log and could we go to row 136.
       So this is your call on 11 May, we see that from
       column D and if we could just scroll right a little way.
       Just pause there.  So column N is what the operator is
       told by you and you are saying:
           "Doing BP ..."
           What's "BP" there?
   A.  Balance period.
   Q.  Balance period:
           "... and got a shortage in cash £17,000.  Had a rem
       in of [46,500] ..."
           There's a typo there:
           "... cash 16,000 coin."
           And it appears from this document that that's what
       you told the operator, is that right?
   A.  I didn't tell them 16,000 coin.
   Q.  Okay.  And then if we carry on to column V this is what
       they told you to do, or suggested that you do:
           "Asked the ..."
           It must be postmaster:
           "... to make sure only one cash declaration and make
       sure rem was scanned in correctly."
           So essentially what you were being told, I think, is
       that you need to check that you have counted the cash
       properly, is that right?
   A.  Yes.
   Q.  Now, if we go back to -- apologies for all the dotting
       around, but if we go back to 1507.1 {F/1507.1} and if we
       could go to row 14515.
           Just while that is being brought up, Mr Patny, your
       evidence in your witness statement is that what the
       helpline told you to do was to readjust the cash stock
       figure; do you remember that?
   A.  Yes.
   Q.  And by that I understand you to mean that you needed to
       carry out a fresh declaration?
   A.  Yes.
   Q.  And you say in your -- or at least I think you say in
       paragraph 10 of your witness statement {E1/2/2} that
       that resolved the problem?
   A.  I believe it did, yes.
   Q.  Okay.  And I just wanted to have a look at 14515 because
       as far as -- if we look at 14515 that's the discrepancy
       on 11 May for £17,000, do you see that?
   A.  Yes.
   Q.  And I can't see anything else after that that is a fresh
       declaration on that day.
   A.  Yes.
   Q.  And I wondered why that was.  I understood from your
       evidence that you had been told to carry out a fresh
       declaration, that you had done so?
   A.  Yes.
   Q.  But that's not recorded there?
   A.  I'm not too sure why.
   MR JUSTICE FRASER:  We saw a document yesterday with two
       cash declarations on, did we not?
   MR HENDERSON:  We did.  We're coming to that one later,
       my Lord.  That's a different date.
           So you can't help me with what happened then, but
       your witness statement suggests that you were told to
   A.  Yes.
   Q.  You did recount and you sorted the problem out.  That
       was my understanding.  Is that a fair understanding?
   A.  I believe I did, yes.
   Q.  Okay.  So what I would have expected to see is on the
       events data that there was a discrepancy, you found the
       money and it was corrected?
   A.  Yes.
   Q.  But that's not shown there.
   A.  I'm not too sure why that's not shown there.  I remember
       we went through quite a few steps, I had to readjust
       various things and then eventually it seemed okay and
       I carried on.
   Q.  Okay.
   MR JUSTICE FRASER:  When you said you found the money, does
       that mean after you did those steps, the physical amount
       of money --
   A.  After I did the physical steps it seemed that everything
       was in order so I carried on.
   MR JUSTICE FRASER:  So there was no longer a discrepancy of
   A.  No.
   MR HENDERSON:  When you say "the physical steps", can I just
       repeat his Lordship's question: do you mean that you
       found the money?
   A.  I didn't find no money as such.  It said -- it seemed to
       me that everything was in order, there was no
       discrepancy, so I carried on.
   Q.  But I mean you are carrying out a physical count of
       money, aren't you?
   A.  Well, after the various adjustments that we did it
       seemed that everything was okay.
   Q.  What adjustments were they then?
   A.  I can't remember.  There were various things that they
       made me go over the phone on.
   MR JUSTICE FRASER:  Could you maybe keep your voice up,
       Mr Patny.
   A.  Sorry.
   MR JUSTICE FRASER:  So whatever steps they told you to do
       over the phone, you did them?
   A.  Yes.
   MR JUSTICE FRASER:  And were they steps to be taken on the
       Horizon terminal?
   A.  It was, yes.
   MR JUSTICE FRASER:  Right, okay.
   MR HENDERSON:  But you didn't make any fresh cash
   A.  I can't remember.
   Q.  Let's look at the cash management report for your
       branch.  Again we looked at this yesterday with your
       father, so I can hopefully take it reasonably quickly.
       In {F/1514.1} -- actually just before -- it might be
       quicker if I take a further reference in this document
       before we leave it, do you mind?  Could we go to 15030
       in this current document, just to speed things up.
   MR JUSTICE FRASER:  Where are we going now?  15030.
   MR HENDERSON:  15030.  And this, Mr Patny, is a cash
       declaration on 13 May, a further declaration, and you
       see in 15029, do you see that you declare cash for
       a little over £80,000?  Do you see that?
   MR HENDERSON:  15029, "Declare cash total 80,000"?
   A.  Yes.
   Q.  And then below that the variance check discrepancy is
       plus 17,000-odd, do you see that?
   A.  Yes.
   Q.  So that appears to cancel out the previous discrepancy,
       doesn't it?
   A.  It does.  Yes.
   Q.  So just pausing at this point, as far as the branch is
       concerned and as far as these records are concerned,
       there is no longer a cash problem, is there?
   A.  Is this saying that it is £17,000 up now, or is it
       saying that --
   Q.  I think that that is saying that whereas before you were
       minus 17, you're now on this one plus 17 so you're
       cancelling each other out?
   A.  Well, if it was cancelling it would be zero, wouldn't
   Q.  Well, I think --
   MR JUSTICE FRASER:  You might want to just take instructions
       to be absolutely clear.
   MR HENDERSON:  I'm told that is right, what I have said.
   MR GREEN:  Sorry, what?
   MR HENDERSON:  So my understanding is that the discrepancy
       that is shown here has the effect of cancelling out the
       previous minus figure, effectively.
   A.  But then that would be zero, wouldn't it, it wouldn't be
       17 --
   Q.  Well, the overall net position would be zero, yes.
   A.  But then that's what that would show there.
   Q.  I'm suggesting that it wouldn't, that what this is doing
       is taking a later comparison on the information that was
       available to it before, which will include the previous
       cash declaration, and is saying that compared to that
       position it's £17,000 up?
   A.  But when you are doing the cash decs every day it tells
       you a running total of whether you are up or down which
       is what you're showing there, so if that in essence
       would mean we are now £17,964 up, which is -- well, that
       means I've found £34,000 all of a sudden from the day
   Q.  My understanding is that's not the right position.  I'm
       sure you can keep a running view of it as well, but that
       what this line demonstrates is as I have previously
           A clearer way of putting the point.  You are 17,000
       down on the 11th, you roll the branch, that's the
       starting position, this is the result of the latest cash
       declaration which is 17,000 up and the net position is
       to balance broadly.
   A.  Okay.  But like I said, the previous cash dec was after
       7 o'clock so as far as I'm aware before 7 o'clock all
       discrepancies or whatever would have been done before
       7 o'clock would have rolled over.  I'm not 100% certain
       on that but that's what --
   MR JUSTICE FRASER:  I'm not sure Mr Henderson necessarily is
       to be honest.
           When you say "roll the branch" would you like to put
       to the witness exactly what you mean by that because we
       had quite a lot of evidence about that before Christmas.
   MR HENDERSON:  Well, I believe the position is that the
       position you have reached on the 11th a rolled
       forward --
   A.  Yes.
   Q.  -- and then when you take the next declaration which we
       are looking here on the 13th -- I can't quite recall
       what happened on the 12th, but anyway on the 13th -- it
       takes that rolled forward position as the starting
       position and it measures it against that?
   A.  Yes.
   MR JUSTICE FRASER:  And your case is that that line in 15030
       means there is no discrepancy, is that correct?
   MR JUSTICE FRASER:  Right.  That's what's being put to you,
       Mr Patny, that as at the 13th there should be no
   A.  Yes.
   MR HENDERSON:  Cash has been found, in other words.
   A.  But like I say the previous day's was done after
       7 o'clock so whatever I declared before 7 o'clock would
       have been taken, so if I had declared zero,
       for instance, before 7 o'clock, that's -- from my
       understanding that's what would have showed the next
       day.  Because the minus 17,000 you have shown me has
       happened at 19.04 which is after the polling time.
   Q.  Let's have a look at what I'm suggesting may be the
       explanation which was the document I was going to go to,
       which was {F/1514.1}.  And this has got some rows
       highlighted.  This is the cash management report for
       your branch on various days and if we could just scroll
       right until we can see the M column, you can see that on
       10 May -- do you see that, which is row 382, do you see
   A.  Yes.
   Q.  There is £16,070 worth of £10 notes, do you see that?
   A.  Yes.
   Q.  And then we know -- I don't need to take you through
       that, but we know that there's £16,000 worth of £10
       notes remmed in on 11 May, do you recall that?
   A.  I believe there was.  I wasn't there when the remming in
   Q.  No, but that's your father's evidence and we have looked
       at that.
           So what you would expect on the 11 May at the end of
       the day is that there would be roughly £32,000 worth of
       £10 notes, less any notes which you had paid out?
   A.  Yes.
   Q.  In fact we can see that there's only 22,130, do you see
   A.  Yes.
   Q.  Which seems strange?
   A.  Slightly, yes.
   Q.  And again I went through this with your father.  We know
       that there's no further £10 notes remmed in on the 12th
       or the 13th, but if you look at those entries, on the
       following day you have gone up by £15,000-odd, do you
       see that?
   A.  Yes.
   Q.  Now, that suggests, doesn't it, that at around this time
       there was some sort of problem with counting cash, but
       it was resolved, you found the cash somewhere?  Perhaps
       it had been put in a safe.  Do you think that's
       a possibility?
   A.  I'm not sure.  It potentially could be, yes.
   Q.  So doing the best that we can at this stage, looking at
       all the records that we've got, it doesn't look as
       though there is a problem with £17,000, does it?  It
       looks as though any problem with the cash has now been
   A.  No, because the cash dec still shows £17,000 in plus,
       doesn't it, so ...
   Q.  Okay, well, we have been through that.
           Let me --
   MR JUSTICE FRASER:  Are you moving off that point?
   MR HENDERSON:  I'm moving off that.
   MR JUSTICE FRASER:  Right.  I just want to be clear about an
       important point that Mr Henderson has been exploring
       with you round the edges and I just want to be
       absolutely clear that I understand what your evidence
   A.  Yes.
   MR JUSTICE FRASER:  When you thought you were £17,000 down
       and you went through the steps on Horizon, did that lead
       to you discovering physical cash of £17,000 --
   A.  No.
   MR JUSTICE FRASER:  -- elsewhere -- just wait for me to
       finish -- which you hadn't counted before?
   A.  No.
   MR JUSTICE FRASER:  It didn't.
   A.  No.
   MR JUSTICE FRASER:  So when you say in different places in
       your evidence "It all looked okay" or "It all worked
       out", what is it that you mean please?
   A.  From what I understand is when I did my --
   MR JUSTICE FRASER:  From what you can remember.
   A.  Yes.  When I did the various adjustments online at the
       end of it there looked like there was no discrepancies
       at all.
   MR JUSTICE FRASER:  In terms of information on a screen.
   A.  Yes, they had made me change various things and that
       made it okay.
   MR JUSTICE FRASER:  All right.  And by "made it okay" you
       mean what?
   A.  The discrepancy wasn't something that -- well, wasn't
       alarming, as such.
   MR JUSTICE FRASER:  Right.  Over to you.
   MR HENDERSON:  My Lord, just to be clear, the Post Office's
       case is that there were various declarations and various
       steps taken and at this period it resulted in no
   MR JUSTICE FRASER:  Well, it depends how you define
       "no problem", Mr Henderson.  The way you were putting it
       seemed to suggest, in at least two questions, that money
       had been discovered physically and that's very important
       and that's why I needed to clarify what his evidence
   MR HENDERSON:  I understand.
   MR JUSTICE FRASER:  Because if you roll them up in terms
       that can be open to misunderstanding the evidence won't
       necessarily be very clear for me.
   MR HENDERSON:  I understand, my Lord.
   MR JUSTICE FRASER:  I know what the Post Office's case is,
       I think.
   MR HENDERSON:  I understand.
           I just wanted to ask you about something on document
       {F/1834.3} and this is a separate point, I think, and
       again I touched on it with your father, but let me just
       ask you.  If you could go to sheet 1 and then row 2336
       and this is the point your Lordship mentioned this
       morning when there are two cash declarations one after
       the other.
   MR JUSTICE FRASER:  And what are we calling this document?
   MR HENDERSON:  This is the events spreadsheet.
   MR JUSTICE FRASER:  The events data spreadsheet?
           Do you see that, Mr Patny?  Row 236?
   A.  Yes.
   Q.  Actually 2336.  So this is 17 May 2016 and there's two
       cash declarations only seven minutes apart but for
       a significant difference, do you see?
   A.  Yes.
   Q.  About £16,000 difference.  Now, I think -- do you accept
       you made these declarations?
   A.  Yes.
   Q.  And again -- and I don't need to turn this up I don't
       think, unless you want me to, but we looked yesterday
       with your father that the transaction data shows that
       all that had happened between those two times was a --
   A.  £300 dec.
   Q.  -- £300 declaration?  Okay.  Can you explain how that
       can happen in that way, that there can be two
       declarations so close to each other for such differing
   A.  From what I can see there that I must have been £16,000
       up for some reason, I have entered in the wrong figure
       and after when I have changed it it has come to the
       second one.  I mean that was -- sometimes when you were
       entering figures sometimes you could enter them in wrong
       and, you know, you would end up with a massive gain or
       a loss and then when you went through and re-entered
       them again it would come to a more correct figure.
   Q.  When you say you re-enter them again, would that involve
       physically counting cash again --
   A.  Sometimes.
   Q.  -- or checking the piles of cash or whatever it is
       you're doing?
   A.  Sometimes, but sometimes it could just genuinely be
       a case of, you know, as you're rushing through pressing
       tab keys -- you have to bear in mind we were serving
       customers whilst we were doing this as well.  It was
       a fairly busy branch after closing times -- well,
       general post office closing times are half past five;
       because we were one of the very few in the city that
       were open after 5.30 so we would get customers coming
       from all around, other side of the city coming to us
       because they knew there was a post office open, so as
       you're rushing through sometimes you didn't tab
       correctly and you would end up instead of writing 20,000
       £20 notes you could add the £10 notes of say 10,000
       £10 notes into that as well so you could end up with
       40,000 in £20 notes.  Just a minor mistake can cause ...
   Q.  Okay.  I just wanted to ask you a few questions about
       the phone call you had from Debra Lambley, do you
       remember this, from Chesterfield about stamps?
   A.  Yes.
   Q.  Again, this was a call that was prompted by a concern
       that you had that you had declared a much higher value
       of stamps than you could conceivably have been holding
       on to, is that right?
   A.  Yes.
   Q.  And that was a declaration that you made?
   A.  I never made the declaration of the stamps.  I don't
       know how they got there.
   Q.  Who made the declaration of stamps then?
   A.  I did the declaration of stamps, but I never put that
       much in.  I don't know what --
   Q.  I'm sorry, I didn't hear.  You said you made the
       declarations --
   A.  I declared the stamps but I never entered in that amount
       on the £1.09 stamps or whatever that was that caused the
   Q.  Well, let's have a look at that this.  Well, just before
       I do that, you were the only one who declared the
   A.  Yes.
   Q.  The figure for the declared stamps was -- we will come
       to it, but it's £18,000.  Everyone agrees it is wrong.
   A.  Yes.
   Q.  Can you explain how that came to be wrong?
   A.  No.
   Q.  Isn't it possible that you just made the sort of error
       that you have just described to me?
   A.  No.
   Q.  Why is that not possible?
   A.  Because you would -- when you entered in the stamps --
       with cash decs, for example, you have to go through all
       of it at the end and then see your variances.  With
       stamps are you literally just enter them in and then
       a report is printed out saying "You have said this much
       is in there", so I would have seen at this point I'd put
       in £18,000 worth of stamps.
   Q.  But it's still possible to make a mistake in that
       process, isn't it?
   A.  Sometimes, yes, if I didn't check it correctly.
   Q.  And isn't it overwhelmingly likely that that's what
       happened here, you made a mistake in declaring the
       stamps which was picked up and corrected?
   A.  Okay, but then the same issue has came back, so
       I couldn't make the same mistake twice, could I?
   Q.  Well, let's look at it.  Let's look at the call log
       first of all at {F/1522.1}, row 138.  If you just go
       right to column N:
           "Call today from Chesterfield and £16,000 adjustment
       now down in the office ..."
           Not quite sure exactly what that means.
       "Activity: declarations and adjustments" and if you go
       right a couple more, you were told to balance the stock
       unit, weren't you?
   A.  Yes.
   Q.  Which is essentially recount, isn't it?
   A.  It was enter in all the figures again, that's what
       I counted it as.
   Q.  That's what I mean, yes.  And if we look at the events
       data spreadsheet for the declaration of stamps, that's
       at {F/1507.1}.  This has been filtered and we can leave
       the filtering on for present purposes because this is
       what we need.
   MR JUSTICE FRASER:  What is it filtered, just to show the
       stamp declaration?
   MR HENDERSON:  The stamp declaration I believe, yes.
   MR JUSTICE FRASER:  Well, that is how it has been filtered?
   MR HENDERSON:  Yes, it is, my Lord.
           So the reason the rows aren't sequential is because
       it has been filtered, Mr Patny, okay?  But if you look
       at the dates you can see that there is reasonably
       regular stamp declarations, generally once a week.  If
       you look at the 11 May one, stamps are declared for
       1,633, do you see that?
   A.  Yes.
   Q.  And then they go up to £18,000, which we all agree is
       wrong I think?
   A.  Yes.
   Q.  And it stays at that sort of level until 25 May, the
       following week, and it goes down to 1,551.  So there has
       obviously been a mistake here, hasn't there?
   A.  Not by me there's not.
   Q.  How can you be so sure it is not by you?  No one is
       saying it's the worst mistake to have made, but there
       has been a declaration of stamps in the branch which has
       been inaccurate.
   A.  Yes.
   Q.  It has been picked up by Chesterfield and business as
       usual, they have said "You need to sort it out", you do
       sort it out and that's what happens.  Why do you say
       that that's a bug in Horizon?
   A.  Because after I've sorted it out, it's stayed there.
   Q.  I'm sorry, I didn't hear?
   A.  After I'd sorted it out, the issue has stayed there,
       hasn't it, it has gone back up to that many stamps up?
       So why would I change it down and then change it back
       again?  I'd understand you said that I'd made the
       mistake the first time, but I couldn't make the mistake
       again the second time, could I?
   Q.  I don't quite understand the point you're making?
   A.  I'm saying you're saying I potentially could have made
       a mistake the first time.  If I did make a mistake the
       first time, fair enough, but I couldn't make the same
       mistake again, could I?
   Q.  This just suggests that stamps are being declared at the
       same sort of level for that week until 25 May and it is
       then corrected, isn't it?
   A.  Yes, but the stamps issue came back.
   Q.  When?
   A.  The following week.  After I adjusted them they still
       went back up.  Debra had to phone me twice.
   MR JUSTICE FRASER:  It might help if you did some of this by
       reference to his witness statement, Mr Henderson.  We
       are having a lot of "I don't understand".
   MR HENDERSON:  I apologise, my Lord.
           Well, I don't understand and I apologise for that if
       it is my error, but are you saying that you carried out
       a declaration which took it down, took the value down
       and it went back up again?
   A.  Well, yes.  Debra -- sorry, I have forgotten her
   Q.  Lambley I think.
   A.  Debra Lambley phoned in the morning, spoke to my father.
       When I came in I spoke to her.  She made me readjust
       them down and what she made me do was very complicated,
       she made me go into areas of Horizon I could never dream
       of going into and then this issue of the extra stamps
       came back after I had done that, so then she's had to
       phone me again the following week and her words were
       "Why have you changed them back?" and I said
       "I haven't".
           So in answer to your question I could have made
       a mistake the first time, okay potentially I could have,
       but then why would I make the same mistake a second
   Q.  So what you have said -- if you look at your witness
       statement in paragraph 13 {E1/2/3}, you say:
           "She told me it was showing the branch had
       overdeclared the amount of stamp stock that was held."
           And we have gone through that?
   A.  Yes.
   Q.  "Ms Lambley told me I had to adjust the stamp stock on
       the system."
           And you say that you did that.
   A.  Yes.
   Q.  All right, that's not shown in that declaration that we
       just looked at?
   A.  But she checked it on her side and said that everything
       was fine as well.
   Q.  And then you say you rang the helpline and said that
       they had given incorrect advice and you had to readjust
       the stamp stock?
   A.  No, when I rang the helpline that night I was £16,000
       short and they made me do various adjustments to make it
       okay again and what this has done is somehow the stamps
       have gone back up and Debra has had to ring me again the
       next week.  So then as I was answering your original
       question, if I made the mistake of entering extra stamps
       the first time, I haven't definitely done it the second
       time, have I?
   Q.  Well, I accept it is unclear, Mr Patny, but what appears
       to have been the case according to the stamp
       declarations is that there was a problem which persisted
       for a week and then was corrected.
   A.  But it hasn't corrected though.
   Q.  Well, it was corrected on 25 May.
   A.  On the day -- Debra rang me the first day, she has
       corrected it.  The issue has persisted, she has rang me
       again a couple of weeks -- well, a week and a half later
       and she has made me do the same thing again.
   Q.  And is it possible that you hadn't properly followed the
       instructions that you were given by her over the phone
       at the time?
   A.  No, because Debra checked it on her side and said that
       everything was okay.
   Q.  But she wouldn't know what stamps you had in the branch,
       would she?  You would have to count it properly?
   A.  Well, if she didn't know what we had in the branch then
       how did she know we were £18,000 up on stamps then?
   Q.  I think the point about that is it is just
       an extraordinarily high amount which everyone accepts
       would not be the number of stamps in a branch of your
   A.  Well, then you can't have it in -- you're talking
       hypothetically in her defence but talking factually in
       mine, aren't you?
   Q.  No, because I think it is agreed, isn't it, that you
       didn't have anything like £18,000 --
   A.  No.
   Q.  -- worth of stamps in your branch?
   A.  No, we didn't.
   Q.  You did have that number?
   A.  We didn't.
   Q.  You didn't, okay.  So what happens is a process whereby
       it is adjusted from a high figure that's been wrongly
       declared to the correct figure?
   A.  Yes.  And Debra has agreed at that point that everything
       is okay.
   Q.  Okay.  Well, I don't think I can take that any further
       for now.
   MR JUSTICE FRASER:  Can I just check what was your
       understanding of "everything is okay", just in terms of
       what's shown and what your understanding was?
   A.  Yes, it is.  Debra has checked on her side that the
       correct amount of stamps are now in holding and
       everything is fine with our branch, but if you have
       a look the stamps clearly wasn't reduced down but she
       has verified that everything was okay and from my side
       I verified everything was okay and I followed what she
   MR HENDERSON:  I want to ask you now, Mr Patny, about the
       MoneyGram transaction.  Do you remember this one?
   A.  Yes.
   Q.  And your complaint here is that you had a MoneyGram
       transaction for money to be sent in the amount of
       £3,100.  That transaction didn't go through, it was
       cancelled and retried, didn't go through again.
   A.  Yes.  The actual transaction wasn't cancelled, just the
       debit card.
   Q.  Okay.  And you say that you had a shortfall of exactly
   A.  Yes.
   Q.  So just to understand what had happened, to make sure
       I have understood this properly, presumably someone
       comes in, wants to carry out a MoneyGram transaction,
       you put it onto the stack on the screen, you then try
       and clear the customer's card and it is declined?
   A.  Yes.
   Q.  But by which stage you have already got it on your stack
       and you have to do something with it in order to balance
       the books, if I can just put it in that way?
   A.  Yes.
   Q.  Okay.  So can we first of all look at the transaction
       data which is at {F/1436.1} and if we go to sheet 1 at
       the bottom please and enable editing and row 1067.
       Thank you.  So this is, I believe, the transaction.  Do
       you see that £3,100?
   A.  Yes, I do.
   Q.  And what we see is a -- we see two appearances of
       £3,100, but that's a single transaction, yes?
   A.  Yes.
   Q.  But what we don't see at around this time is a second
       entry for £3,100.  We do not see that.  Your evidence
       effectively is that we should see that because you say
       that this caused a problem of £6,200.
   A.  No, but from -- what we used to do with MoneyGram, the
       procedure was you cash it through, so you cash -- say it
       has gone through and then you cancel it at that point.
       I have done that correctly, but the Post Office had
       changed the procedure in December which I wasn't aware
       of that you had to cancel it on the back office as well
       at that point, so I've never cancelled it there, so that
       was still hanging.
   Q.  Okay.  That's what I wanted to come to.  So you are
       absolutely right, there is a two-stage process.  Could
       I just hand this document to the -- this is the document
       I handed up to your Lordship earlier.  Apologies,
       Mr Patny, for the paper.  This is finding its way into
       the electronic form.
           This is the procedure which was online and this is
       an operational update, as you will see, from
       19 October 2015.  Do you see that date in the top left?
   A.  Yes.
   Q.  And this is the procedure for dealing with the situation
       that had arisen and if you look at page 4, and you have
       just described it I think, it says:
           "The refund cancellation process for MoneyGram has
       changed ..."
   MR JUSTICE FRASER:  Whereabouts are we, sorry?
   MR HENDERSON:  I'm so sorry, my Lord, it's the final page of
       this clip, the fourth page.  It says "Page 4 of 4" at
       the top.
   MR JUSTICE FRASER:  It says "Page 4" on the top?
   MR HENDERSON:  4 of 4 on the top left, yes.
           It says:
           "If a customer wants to cancel their transaction and
       obtain a refund on the same day you must go through the
       MoneyGram refund button to cancel the transaction.  You
       then need to use the back office reversals process to
       reverse out the transaction."
           And that's the point you are just making, isn't it?
   A.  Yes.
   Q.  So you were supposed to follow a two-stage process which
       is cancellation and reversal?
   A.  Yes.
   Q.  And you just did cancellation --
   A.  Yes.
   Q.  -- not reversal?
   A.  No.
   MR JUSTICE FRASER:  "No" as in you agree you didn't do
   A.  Yes, I agree with what ...
   MR HENDERSON:  And that procedure, that was part of the
       Horizon Online guidance at the time.
   MR GREEN:  Well, I'm sorry, my Lord, it should be pointed
       out to the witness the date of the document that's not
       on Opus yet but we have just seen.  The date of this
       document is "Operational updates of 19 October 2015", so
       it is fair to -- because you can put something in front
       of someone and it may influence their confidence about
       what they can remember.
   MR HENDERSON:  Okay.  In any event you now accept that there
       was a two-stage process --
   A.  Yes.
   Q.  -- that in an ideal world you would have followed, which
       is cancellation and reversal.
   A.  (Nods).
   Q.  And you rang the help desk, didn't you, in relation to
   A.  I did, yes.
   Q.  Let's have a look at {F/1522.1}.  If we go to row 115.
   MR JUSTICE FRASER:  Is this the call log?
   MR HENDERSON:  Yes.  This is your call -- do you see this,
       Mr Patny?  This is 23 February 2016, at column D, and if
       you go across and then just pause there at N, column N,
       you said:
           "Debit card declined for MoneyGram, how to clear."
           So you had rung to ask what to do about this?
   A.  Yes.
   Q.  And then you go across and the advice was "settle to
       cash and then cancel and reverse".
   A.  Yes.
   Q.  So you were told you did need to reverse it but for
       whatever reason you didn't understand that or you didn't
       do it, but in any event you didn't reverse it?
   A.  I didn't reverse it at the time of the transaction but
       I reversed it when the helpline helped me through it.
   Q.  Well, according to Post Office it was never reversed.
       It was only cancelled, not reversed.
   A.  Sorry, are you on about the 7 o'clock phone call here or
       the 12 o'clock phone call?
   MR JUSTICE FRASER:  This phone call is -- well, let's go
       back to the left-hand side and we will see the time.
       Could we go back to column A.  This is at 12.50.
   A.  Sorry, I was talking about when I rang them later on in
       the day.
   MR HENDERSON:  You rang them later on in the day?
   A.  Yes.
   Q.  Well, I had understood you to say before in your
       evidence that you accepted that you had not reversed the
   A.  No, I hadn't reversed it in the back office, no.
   Q.  Are you saying that you did reverse it in due course?
   A.  I did after I spoke to the help desk, who they told me
       I had to reverse it at the back office.
   Q.  Okay.  So as far as the data is concerned and as far as
       the Post Office is concerned you did not reverse it.
       You did cancel it.
   A.  Yes.
   Q.  But you did no not reverse it?
   A.  No.
   MR JUSTICE FRASER:  At 12.50.
   MR HENDERSON:  At all.
   A.  As the transaction took place I believe at midday,
       I cancelled it, I never reversed it at that point.
       I reversed it in the evening when I did my cash dec and
       saw I was short and I reversed it after speaking to the
       NSBC[sic] because at that point I was still unaware that
       I had to reverse it on the back office.  Reversals take
       place on the back office, cancellations was on the front
   Q.  My suggestion to you, Mr Patny, is whether or not you
       understood the process, as a matter of fact you didn't
       effectively reverse the transaction?
   A.  I didn't reverse it when I should have, but I did later
   Q.  Okay.  I'm just being clear, I am saying you didn't
       succeed in reversing the transaction at all?
   A.  But we did, we reversed it later on, didn't we?
   Q.  Well, as far as the evidence the Post Office can see
       that didn't happen.  It did cancel, didn't reverse?
   A.  It did cancel and it did reverse but because that's
       after -- once again, after 7 o'clock, the issue was the
       day's data had been polled so I have -- effectively it
       doesn't look like I have done it on that day, where in
       terms of date I have but for business dates I haven't
       done it on that date.
   MR JUSTICE FRASER:  And you think there's a second phone
       call, did you say?
   A.  Yes.  That's here, this quarter past 7 one, that's the
       second phone call.
   MR JUSTICE FRASER:  When you say "that's here", what are you
       pointing out?
   A.  116 is it?  No, sorry I saw it -- if you scroll across.
       There, 116, "Cash dec £6,000 short due to MoneyGram".
       That took place at quarter past 7.
   MR HENDERSON:  Okay and you say "Short due to MoneyGram
       refund did not do the existing reversal".
   A.  Yes.
   Q.  Okay.  And carry on across.  Just look at 117 below
       that, what does that involve, "Cancellation refund"?
       Can we just go back to the beginning of 117.  And can
       you go across from 117.  Here you are ringing again
           "Asking if the discrepancy has been resolved yet -
       waiting for a TC to be issued - had failed to reverse
       a [MoneyGram] for £6,000 that was cancelled on the same
       day as having been sent."
   A.  So the next day when I rang up I have asked "Where's
       the TC?"; I wasn't aware that a TC took several days to
   Q.  Okay, but there is a single transaction here of £3,100?
   A.  Yes.
   Q.  Right.  The process is it needs to be cancelled and
   A.  Yes.
   Q.  You only cancelled it?
   A.  Yes.
   Q.  You then think you may have reversed it?
   A.  I did reverse it, as you can see there, quarter past 7
       in the evening.
   Q.  Well, on 117 it says that it had failed to reverse
       a MoneyGram for 6,000, which is not right, it's
       a MoneyGram for 3,100?
   A.  Yes.
   Q.  And can we carry on going across on 117.  Okay, so doing
       the best we can, Mr Patny, and I appreciate these things
       are not as straightforward as they could be, it looks as
       though there has been a transaction for £3,100 --
   A.  Yes.
   Q.  -- which has not been reversed properly?
   A.  At the time --
   Q.  In other words you need a transaction correction for
   A.  We do, yes.
   Q.  And just to be quick about it, you in fact receive
       a transaction correction for £3,100, don't you?
   A.  We do, several days later, yes.
   Q.  Yes, several days later.  But your complaint is that you
       were £6,200 out-of-pocket, is that right?
   A.  Yes.
   Q.  And what you say in paragraph 20 of your first witness
       statement {E1/2/4}.  What you have said in paragraph 20
       is that you thought it was strange that it was exactly
       double the amount of the problematic MoneyGram
       transaction.  Do you see that?
   A.  Yes.
   MR HENDERSON:  My Lord, I just wondered -- this is going to
       take a little while.  I wonder if it might be sensible
       to have a brief break now?
   MR JUSTICE FRASER:  Very good idea.
           We're going to have a short break.  Today it is
       going to be seven minutes long.  It is for the shorthand
       writers just to have a break from transcribing
       everything word for word.  Because you are in the middle
       of your evidence you're not allowed to talk to anyone
       about the case but you don't have to stay in the witness
       box.  Feel free to move, go out, stretch your legs.
           If everyone could be back in at quarter to.
   (11.38 am)
                          (Short Break)
   (11.46 am)
   MR HENDERSON:  Mr Patny, I want to just ask you a couple
       more questions about the stamp situation before I come
       back to the MoneyGram situation that we were dealing
       with before.
           Do you agree with this, that if a declaration of
       stamps goes up, the cash position is automatically
       decreased, isn't it?
   A.  I'm not sure.  I don't know that.  As in do you mean for
       overall, or holding --
   Q.  It automatically happens, doesn't it, that as the
       declaration of stamps goes up, the cash position will go
   A.  I don't see why that would.  The cash is going to remain
       the same, isn't it, if stamps go up?
   Q.  Let me put it in a different way.  If you were trying to
       disguise a shortfall of cash, one thing you could do is
       to declare too much by way of stamps, isn't it?
   A.  No.
   Q.  You disagree with that?
   A.  Yes, because why would stock make a difference to cash?
   Q.  Because the more stock you've got, the less cash will
       show, as a matter of balance?
   A.  Potentially, but I'm not too sure of that.
   Q.  Are you confirming that all of the declarations that you
       made in relation to both cash and stamps at all times
       were accurate to the best of your knowledge?
   A.  I tried to, yes.
   Q.  And a declaration of stamps -- you remember the £18,000
       declaration of stamps, was that an attempt by you to
       disguise a shortfall of cash?
   A.  No.
   Q.  I suggest it may have been.  I suggest it was, Mr Patny.
       You deny that?
   A.  I do, yes.
   Q.  I suggest finally, Mr Patny, it is rather surprising,
       given that you do all the balancing, or did all the
       balancing in the branch, that you don't understand that
       the effect of increasing the declaration of stamps is to
       have a corresponding effect on the declaration of cash?
   A.  I wasn't aware of that, no.
   Q.  All right.  So let's go back to the MoneyGram
       transaction and just to remind you where we had got to
       with that, there's a transaction for £3,100, you recall
   A.  Yes.
   Q.  And you say you reversed it effectively?
   A.  At that point, or ...?
   Q.  Well, my understanding is that your evidence was that
       you did reverse it properly?
   A.  Eventually I did, yes.
   Q.  Separately from the transaction correction?
   A.  Yes.
   Q.  So you both reversed it and you received a transaction
   A.  Yes.
   Q.  That makes you £3,100 up, doesn't it?
   A.  No, because the reversal happened on the wrong
       day --well, on the wrong business day, because once
       again if you have a look at the call log, I spoke to
       them at quarter past 7 and they told me to do this,
       7 o'clock is polling.  So it would have polled as the
       £6,200 short for the day.
   Q.  Mr Patny, it has nothing to do with the day.  If you
       have reversed the transaction with the effect that
       you're no longer losing that £3,100 and you're getting
       a transaction correction for £3,100, you end up £3,100
       up, don't you?
   A.  No.
   Q.  Is that what you're saying happened?
   A.  No, we were -- would it be easier if I explained from
       the start what's happened with this MoneyGram issue?
       Would that make it easier for my Lord, or ...?
   Q.  By all means.
   MR JUSTICE FRASER:  Go ahead.
   A.  A gentleman came in.  His MoneyGram has declined for
       3,100.  His card has declined twice.  I have cancelled
       it at that point on the front office but I haven't done
       anything on the back office which is what I should have
       done, but I wasn't aware of that because it had changed
       a couple of months earlier.  Later on that night I'm
       £6,200 short.  I phoned up the helpline to ask why this
       was and immediately "Look 6,200, you've got a £3,100
       short MoneyGram", that to me meant it doubled up
       somewhere and overall there had been an issue there.
       MoneyGram told me to reverse it on the back office, they
       made me do various things and they said --
   MR JUSTICE FRASER:  When you say MoneyGram told you?
   A.  Not MoneyGram, sorry, the helpline told me.  They also
       said a transaction correction would be sent to balance
       out the losses.
           So that's what they made me do.  Now, the issue is
       on that day it would have polled as a £6,200 shortfall
       because at 7 o'clock that's when the polling happens and
       the business day ceases as such, so I finished the day
       as technically the business day I have not corrected it
       on the right day, but date-wise I have actually done it
       on that day, because I have done it after 7 o'clock.  So
       the next day I phoned up to ask -- because we had never
       had an issue like this before of where the TC was and
       where the money was and they said it takes a few days to
       generate and that's what's happened.
   MR JUSTICE FRASER:  Just two questions from me about
       MoneyGram so I can understand.  MoneyGram is where
       somebody wants to send some money overseas.
   A.  Yes.
   MR JUSTICE FRASER:  And it is done in a secure way.
   A.  Yes.
   MR JUSTICE FRASER:  So they come into your post office and
       either give you cash or give you a debit card and you
       take the money from them.
   A.  Yes.
   MR JUSTICE FRASER:  And then the recipient in another
       country can use that MoneyGram account to withdraw that
       money, is that correct?
   A.  Yes.
   MR JUSTICE FRASER:  So when the gentleman came in to try and
       do this transfer and his debit card was declined, am
       I right in concluding that that means you didn't take
       any money off him for his MoneyGram, is that right?
   A.  No, his card declined twice.
   MR JUSTICE FRASER:  Which meant he could not then -- or
       whoever it was, his relative or friend in another
       country could not have the benefit of the money that you
       hadn't taken --
   A.  Yes.
   MR JUSTICE FRASER:  -- is that right?
   A.  Yes.
   MR JUSTICE FRASER:  Over to you, Mr Henderson.
   MR HENDERSON:  My Lord and just to be clear as to the
       Post Office's position which I anticipate your Lordship
       understands, it is accepted that the issue arose in
       relation to 3,100 and there was a transaction correction
       in relation to it, which is accepted --
   MR JUSTICE FRASER:  The issue being ..?
   MR HENDERSON:  The issue is whether there was a double
       transaction for 6,200.  And that's what --
   MR JUSTICE FRASER:  Because you did -- as I understand it,
       and you can correct me if I'm wrong and Mr Henderson you
       can correct me if I'm wrong -- you did get a transaction
       correction for £3,100.
   A.  Yes.
   MR JUSTICE FRASER:  And that took a number of days, is that
   A.  Yes.
   MR JUSTICE FRASER:  Am I right?
   MR HENDERSON:  Correct.
   MR JUSTICE FRASER:  Over to you.
   MR HENDERSON:  But what I have been exploring with the
       witness is his suggestion, which I didn't understand
       from his witness evidence, that he had additionally
       reversed the transaction and if he had effectively
       reversed the transaction he wouldn't be entitled to
       a transaction correction.  But there is a single
       transaction here and the data shows a single transaction
       of £3,100 --
   MR JUSTICE FRASER:  When you say "the data", are you talking
       about --
   MR HENDERSON:  The transaction data.
   MR JUSTICE FRASER:  But not the call log.
   MR HENDERSON:  Not the call log, no, but the transaction
       data.  But the call log suggests, as we were looking at
       before -- it might be sensible --
   MR JUSTICE FRASER:  Well, let's not go into the arena of you
       explaining it to me because you are still in the middle
       of cross-examining him.
   MR HENDERSON:  No, indeed.
           So let's explore the suggestion that I have put to
       you just before the break, Mr Patny, that your loss was
       exactly £6,200.  Do you remember that?
   A.  Yes.
   Q.  So can we look at {F/1507.1}.  If we go to -- you need
       to take the filter off and go to row 1903.  Now, this is
       a bit fiddly, Mr Patny, and I apologise for that but we
       need just to take this in stages.
           So the actual discrepancy that you experienced and
       declared at the end of 23 February was not for £6,200,
       it was for £6,825.95.  Do you see that?
   A.  Yes.
   Q.  So in itself that does not suggest that there is
       a problem for £6,200, does it, in itself?  I'm going to
       go on.
   A.  No.
   Q.  But we need to look at the previous day and if we look
       at -- it may be sensible if you just make a note --
       your Lordship may take the arithmetic from me, but we
       have a declaration of 25,803.87.  Let's look at the
       previous declaration of cash --
   MR JUSTICE FRASER:  Sorry, where?  Oh, you are taking it
       from the line above, 25,803, yes.
   MR HENDERSON:  My Lord, yes, 25,803.87, discrepancy of
       6,825.95.  And if you look at the previous day which is
       row 1717, that is cash declaration for that day of
       34,405.46.  Do you see that?
   A.  Yes.
   Q.  So there is movement between the 22nd and the 23rd in
       the cash declaration figure of £8,601.59.  Perhaps you
       would accept that from me.  Your maths is probably
       better but that's fine.
           So what I want to explore is whether we can see the
       figure of 6,200 somewhere in that figure.  Do you
   A.  Yes.
   Q.  Now, the first thing is we can have a look at what the
       session data shows us for cash transactions over the
       relevant period and this is to be found in document
       {F/1436.1.1}.  Sheet 1.
           My Lord, I should just explain this, and Mr Patny
       I should just explain this document.  This is
       transaction data filtered for trading date and cash
       between the times of the two cash declarations that
       I have just taken you to, Mr Patny.  Do you understand
   A.  Yes.
   Q.  Now, this is a document that we have created for the
       purposes of today in order to demonstrate this point
       because otherwise it is just too painful to make it.
       I have explained it to Mr Green and Mr Green -- if you
       just go to the top of that document -- sorry, before you
       do so, just look at the figure of 1,806.71 which is in
       row 216.  That's the total of the decrease in cash in
       the branch, okay?
   A.  Okay.
   Q.  And if we just go to the top of the document, it is
       right to say that column U, "Calculation", is one that
       we have added, all right?  That's not one that appears
       in the actual transaction log.  So we have filtered this
       in order to demonstrate the movement of cash of
       1,806.71.  Does that make sense?
   A.  Yes.
   Q.  It is a figure that you will see in Ms van den Bogerd's
       witness statement in I think paragraph 74 {E2/5/20},
       I think from memory.
           So taking it in stages, we started with a figure of
       1,806 as the movement in cash.  We've got the MoneyGram
       transaction for 3,100 which we know about which at this
       stage the transaction correction hasn't been issued,
       okay, so that's another 3,100.  So the total decrease
       that we can account for is 3,100 plus 1,806.71, which
       leaves £3,694.88 of cash movement unaccounted for.
           Do you agree with the approach I have taken at
       least?  I appreciate it's very difficult for you to
       agree with specific figures.
   A.  I'm trying to understand what you're saying, yes.
   Q.  Okay.  The point that I'm putting, clumsily and at great
       length, is that there is not a discrepancy of £3,100.
       It is a figure of over £3,600.
   A.  Right.
   Q.  So when you say that there was a shortfall of exactly
       £6,200, that's just not right, is it?
   A.  Well, according to your calculations there it's
       different, yes, but that 6,200 -- when I'm saying
       "exactly" it could be to the pence but I'm claiming that
       3,100 times two which showed up as a shortfall on my
       cash dec.
   Q.  Well, Ms van den Bogerd suggests that it is much more
       likely to have been -- I mean the 3,100 for the
       MoneyGram is one thing, but for the rest of it it's much
       more likely to be a cash handling problem or a user
       error when making cash declarations.  Can you say that
       she is wrong in saying that?
   A.  Well, yes, I can.
   Q.  On what basis?
   A.  On what basis can she say that that's a cash user ...
   Q.  Because there's no indication in any of this data that
       there were two lots of £3,100 that were processed in any
       way.  There is every indication that there was one lot
       of £3,100 and the story that is told in relation to that
       £3,100 makes entire sense.  There is no indication of
       the £6,200.
   A.  The only points I can raise to that is we spoke to our
       area manager, Mark Irwin(?), we showed him all this
       evidence and he agreed with us that it seemed like this
       had doubled up and he is the one who put the block on
       our behalf of Chesterfield asking for this money.  So
       Mark Irwin is obviously a Post Office employee and he
       said this and done this as well.
           Then the second thing, I don't know if I'm being
       obtuse here, but we have requested this data many times
       from the Post Office and they have never gave it once.
   MR JUSTICE FRASER:  The data that you are being shown.
   A.  Yes, I believe this is what's the Credence report --
       I might be wrong, but I believe this is the Credence
       report and we have requested on at least five to six
       occasions of this so we could try to go through it
       ourselves on the behest of Mark Irwin.  He is the one
       who told us to ask for this, as it Tony Sanghera, as did
       the NBSC as well and they said this is where you could
       work out and see if there are some difficulties with you
       guys ...
   MR HENDERSON:  Mr Patny, I'm in some difficulty because
       I can only go on what you've said in your witness
       statement and you haven't dealt with any of those
       matters in your witness statement and I can only go on
       the data that I can show you and again my suggestion is
       there is no possible basis for you asking for a further
       £3,100 when you have already had the transaction
       correction arising out of this transaction.
   A.  Without accusing Ms van den Bogerd, you said yourself
       she has created this so we don't know what she has
       created, do we?  She could have done anything on it.
   MR HENDERSON:  My Lord, I have no further questions.
   MR JUSTICE FRASER:  Re-examination, Mr Green?
                    Re-examination by MR GREEN
   MR GREEN:  Just a couple of points, if I may.
           Could you please look at the helpline logs which are
       at {F/1509.1} and could we please select A and B columns
       just for the moment.  Don't worry, we will just do it
       manually.  We will do it manually.
           Let's go down please to row 140 and could you just
       bring that a little bit higher up please.
           Now, you will see there that's 26 May.  Can you see
   A.  Yes.
   Q.  And if we go across please you can see it is recorded
       there 26 May, 12.33 the record is created, that's the
       "Created" column.  If we go across please you will see
       row 140 still, you see "Aakash" with one A but that is
       you though.  You actually have two As?
   A.  Yes.
   Q.  And then "17,000 shortage open postage".  Now, could the
       operator very kindly click on "£17,000 shortage open
       postage", that cell, and then please click "Wrap text"
       if we can see that.  You might have to go -- I think you
       might have to go to the right top corner.
   MR JUSTICE FRASER:  Do you want to show all the text in that
   MR GREEN:  Yes.  Actually we can go to the top right, little
       down arrow.
   MR JUSTICE FRASER:  I think what you need to do is if you go
       to the very top where the letters for the columns are.
           Is that what you wanted?
   MR GREEN:  I'm very grateful.
           Now, let's just look at the entry on the 26th.  You
       will see there in row 140 "£17,000 shortage open
       postage" and then a reference to 09/05 "£17,000 loss
           Can we please go across one column to the right
       please.  Just scroll to the right one column very
       kindly.  And you will see Horizon Online "Stock",
       "Declarations and adjustments" and then could you please
       click on "Emailed ESG" and then perhaps go up to the top
       and do the same exercise.
           It may be actually easier -- do you see in the
       formula bar at the top just underneath the ribbon
       there's a right little arrow pointing downwards, could
       you click on that arrow very kindly in the formula bar.
       You can just see the text there.
           So this is the entry on helpline log, Mr Patny:
           "Emailed ESG Thursday 26/05/2016 14.22 requesting
       Credence report.  Please see attached documents advised
       office of Credence info.  Office adamant stamps are
       declaring themselves and overriding [IS] figure referred
       to IT if believes system issue."
           Now, you mentioned that you had requested a Credence
       report.  Can you remember whether they told you they
       were providing you with one or not?
   A.  They said they would but we never received it.
   Q.  Okay, well, let's follow that through.  If you then go
       down please, let's go down please six rows to row 147.
       Can we go to the left now please.  If you look at 147
       could we please click on the cell "Office state" in
       row 147.  You can see up at the top we can see what that
           "Office state they were advised a Credence report
       was being submitted for amount £17,200 but no report has
       been requested ..."
           Even though the prior entry suggests that an email
       was sent:
           "... see previous references.  Advised will request
       report and ring office back 22/06 from 11/05, will then
       require an audit if cannot be found.  No details on
       previous calls.  Need to call office back for further
           Then if we go further down in that, so the bit you
       kindly did, can you scroll down a tiny bit there, in the
       formula bar.  Then:
           "Have emailed ESG.  The office are adjusting stamps
       by a similar amount to the discrepancy.  ESG are
       investigating, called office ... spoke to Akash.  I have
       also spoken to Sarah H who has been investigating the
       office ..."
           Can you go down a bit more.
   MR JUSTICE FRASER:  I'm sorry, I can't see where you are
       reading from.
   MR GREEN:  In the formula bar, my Lord, at the top.
   MR JUSTICE FRASER:  Which row is this?
   MR GREEN:  The row that's highlighted is 147.
   MR JUSTICE FRASER:  There are about eight highlighted which
       is why I'm asking.
   MR GREEN:  Sorry, the row that's highlighted with the blue
       round it.  It is row 147 and it is "Office state", the
       bit your Lordship pointed out is in block capitals.
   MR JUSTICE FRASER:  So you are reading from the formula bar
       which is a more full entry.
   MR GREEN:  It is a full entry.
   MR JUSTICE FRASER:  Understood, okay.
   MR GREEN:  So it says:
           "I have also spoken to Sarah H who has been
       investigating this office activity through Horace.
       There is no conclusive information relating to the loss
       that has been found at this time.  However the office
       transactions are still being investigated and the office
       hopefully should be contacted before the end of next
       week 08/07.  PM has been advised this information."
           Now, at this stage what did you understand was
   A.  That they were going to be sorting out all these issues
       that were prior really.  They would sort out a Credence
       report and we would be told what was happening.
   Q.  Okay.  Can we go down two further to 149, which has not
       been highlighted.  And you can see "Previous call".  Can
       you click on that very kindly please.  And you will see:
           "Previous call ... still waiting to hear back
       regarding a requested Credence report."
           Then can we go down please to 3 August which is,
       I will just tell you now ... if we go down to 3 August
       on the left-hand side.  It is row 155.  So if you
       actually just go just to 155 there.  It is not
       highlighted.  It says "ACKAH" is the person.
   A.  I assume that's me.
   Q.  But you assume -- there's no person called Ackah in the
   A.  No.
   Q.  Can you just very kindly click on the cell starting
       1895075 and then it says:
           "Branch still waiting a call back.  Call closed,
       Sarah ESG advised to assign to your stack to call the
       office back.  No details listed on the call regarding
       how the incorrect cash rem is dealt with."
           Did you ever receive back the Credence report you
       understood you were going to get?
   A.  No.
   Q.  And did the sums that we have been looking at here,
       stamps and cash, £17,000, feature in the special audit?
   A.  They did, yes.
   MR GREEN:  No further questions, my Lord.
                 Questions from MR JUSTICE FRASER
   MR JUSTICE FRASER:  I have two questions.
           Can we go to {F/1507.1} please.  Right, that doesn't
       appear to be the document I was expecting it to be.
       Mr Henderson, you are going to have to help me with this
       then because it was the document that showed the
       discrepancies, do you remember?  Row 1903 had the
       discrepancy of 6,825 on 23 February 2016 and a total
       cash figure of 25,803.87 and you were comparing that
       with --
   MR HENDERSON:  Yes, my Lord, that's 1507.1.  If we take the
       filter off.
   MR JUSTICE FRASER:  Oh, I do beg your pardon.  Could we take
       the filter off.  Thank you very much.  Right, if we can
       go to row 1903 please.
           You were asked, Mr Patny, about the discrepancy on
       23 February.  Do you remember those questions?
   A.  Yes.
   MR JUSTICE FRASER:  And you were shown row 1903 which had
       a discrepancy of £6,825.  It had a declare to cash total
       of 25,803 and a discrepancy of 6,825.
   A.  Yes.
   MR JUSTICE FRASER:  And as I understand your answers, the
       £6,200 is within or should be within the 6,825
       discrepancy, is that correct?
   A.  Yes.
   MR JUSTICE FRASER:  Is that a correct understanding?
   A.  Yes.
   MR JUSTICE FRASER:  If you go to the previous day which is
       on 1717, you were asked about the cash declaration of
       34,405.  And that was the cash total I think for the
       previous day, is that right?
   A.  Yes.
   MR JUSTICE FRASER:  Which is 34,405.  And you were then
       taken to another document which I'm not going to take
       you to, to examine the movement of cash from one day to
       the next.
   A.  Yes.
   MR JUSTICE FRASER:  And that I think is the document which
       you said you had been asking for which you referred to
       as a Credence report.
   A.  I believe that's what the Credence report is.
   MR JUSTICE FRASER:  But if you look on 22 February there are
       also two other entries.  One is immediately above 1717
       which shows a cash discrepancy I think of £781 and one
       is a cash discrepancy in row 1718 which shows £31.  Do
       you see those two figures?
   A.  Yes.
   MR JUSTICE FRASER:  If I want to go away and compare like
       with like, should I be taking just the cash declaration
       the day before, or should I taking the cash declaration
       and the discrepancy the day before --
   A.  I'm --
   MR JUSTICE FRASER:  -- and comparing it with the cash
       declaration and the discrepancy for the following day?
   A.  I think the second one would be more beneficial to you,
       wouldn't it, because then you can see --
   MR JUSTICE FRASER:  So therefore is it correct that as at
       22 February whatever the discrepancy was on these
       figures, that is not included in the 34,405 figure; is
       that right?  Or is it?
   A.  That £31.07 -- that 34,405 should be, to my
       understanding, after that £31 discrepancy.
   MR JUSTICE FRASER:  That was my understanding.  And does the
       same point apply for the two figures the day before as
   A.  Yes.
   MR JUSTICE FRASER:  So if ever you look at the cash figure
       you have to look at the cash figure plus the
       discrepancy, is that right?
   A.  Yes.
   MR JUSTICE FRASER:  Thank you very much.
           That's all my questions.  Thank you very much for
       coming.  You are now free to leave the witness box.
           Right, Mr Green.
   MR GREEN:  My Lord, I am calling Mrs Burke now.
                     MRS ANGELA BURKE (sworn)
   MR JUSTICE FRASER:  Have a seat.
   A.  Thank you.
                 Examination-in-chief by MR GREEN
   MR GREEN:  Mrs Burke, there will be a folder in front of
       you.  Move that sheet of paper away.  You can put the
       loose sheet away somewhere else because that was
       separate and if you very kindly have a look in tab 4
       {E1/4/1}, you will see "Amended witness statement of
       Angela Burke"?
   A.  Mm-hm, that's right.
   Q.  And at paragraph 5 of that {E1/4/2} you deal with your
   A.  That's right.
   Q.  It is right that you are no longer working as
       a subpostmistress?
   A.  No.
   Q.  And subject to that correction can we have a look at the
       back page --
   A.  I don't work in the greetings card and stationary
       business either now, that's closed.
   Q.  So you are no longer working there either.
   A.  No.  I've got no employment.
   Q.  And if you look at the final page of your witness
       statement, which is page 6 {E1/4/6}.
   A.  Yes.
   Q.  You see a signature there.
   A.  Mm-hm.
   Q.  Is that your signature?
   A.  That's my signature, that's correct.
   Q.  And subject to those points about no longer working as
       a subpostmistress or in the branch, are the contents of
       your statement true?
   A.  That's correct, yes.  Everything else is fine.
   Q.  Thank you very much.  Would you wait -- sorry?
   A.  Yes, everything else is fine.
   Q.  Would you wait there, there will be some questions from
       Mr Draper.
                  Cross-examination by MR DRAPER
   A.  Hello.
   MR DRAPER:  Good afternoon, Mrs Burke.  You just said you
       were an assistant at Newport which was your husband's
   A.  That's correct, yes.
   Q.  I understand that before that you were also yourself
       a subpostmistress, is that right?
   A.  Yes, that's right.
   Q.  In total then how long have you worked as either
       a subpostmistress or an assistant?
   A.  I have worked with the Post Office since I was 16 on and
       off, sometimes in sub-post offices, sometimes I had
       three -- probably five years as subpostmaster and then
       intermittently as an assistant.
   Q.  Can you give any indication of the number of years
       during which you were an assistant; are we talking five,
       ten or more?
   A.  Let me just think.  Probably about 15.
   Q.  I just want to ask you quickly about a change that you
       recently made to your witness statement.  If you have it
       there in front of you, in paragraph 8 {E1/4/2} you say
       that you and your husband never had any major issues
       with unexplained shortfalls during your time running the
   A.  That's right.
   Q.  You then say there was £1 here or there but never
       anything more than that.  In that paragraph the word
       "unexplained" was added recently.  I'm going to ask you
       what you mean by that but I'm going to suggest what it
       may be.  Is that clarification because you did have some
       shortfalls that were over £1 but you were able to
       investigate and resolve them?
   A.  That's correct, yes.
   Q.  And do you say you were able to do that typically at the
       end of the day when you carried out your cash
       declaration, you would identify the shortfall and then
       look into it, is that right?
   A.  Usually, yes.
   Q.  If you identified a shortfall of a fairly substantial
       size, say £50 or £100, much more than a pound, and you
       were unable that evening to identify where it had come
       from, was your practice to contact Post Office?
   A.  No.
   Q.  Did that situation not arise?
   A.  No -- normally if you had a shortfall and -- you would
       try and investigate it yourself and try and find out
       what the cause of it was and if you didn't know it, you
       would hopefully -- I mean in days gone by it could have
       been that it could be a transaction that had been sent
       off and you had recorded it wrong and an error notice --
       in the old days -- would come.  They have now been
       replaced by transaction corrections.
   Q.  So you're saying sometimes where even if you had
       a relatively substantial shortfall, for the time being
       you might assume that a transaction correction would
       come and correct it?
   A.  That's right, yes.
   Q.  Is that right?  And in your experience is that what
       happened in those circumstances where you had those kind
       of shortfalls?
   A.  Usually, yes.
   Q.  In your witness statement you describe a problem you had
       with recovering transactions on 9 May 2016.
   A.  Right.
   Q.  You give the details of a number of transactions that
       either didn't go through at all or had to be recovered.
       The only one I want to ask you about is the £150
   A.  Okay.
   Q.  Before I do that it would be helpful to take you to some
       images of the Horizon screen so we can make sure that
       we're talking about the same thing when I describe the
       process.  So if we could call up please {E2/4/6} and
       could we zoom in please on the first image on that page.
       This is obviously not a photograph, it is an image to
       represent the front office home screen on
       Horizon Online.  If you look towards the right of the
       screen there, Mrs Burke, where it says British Gas in
       blue -- do you see that?
   A.  Yes.
   Q.  And then below it in red "Total due from customer"?
   A.  Yes.
   Q.  That section of the screen I suggest is what we will
       refer to as the basket or stack?
   A.  Yes, that's correct.
   Q.  And products and services that you enter into the system
       by using the buttons that we see on the rest of this
       image then stack up in that section of the screen, is
       that right?
   A.  Yes, that's correct.
   Q.  And each one of those items when it goes into the stack
       will show a price which may be negative or it might be
       positive and that depends on whether that product or
       service involves payment to or payment from the branch.
   A.  Yes.
   Q.  That's right, isn't it?
   A.  Yes.
   Q.  Could you speak up please.
   A.  Yes, that's right.
   Q.  And underneath the stack, as we have just established,
       in this image it says "Total due from customer" and it
       is in red, but it might also, if the transactions were
       different, say "Total due to customer" in which case it
       would be green?
   A.  Yes, that's correct.
   Q.  And whether you get a green or a red section there
       depends on the net effect of all the transactions in the
       stack, doesn't it?
   A.  Yes.
   Q.  So if we take, for example, this British Gas payment, if
       the same customer in this transaction were then to ask
       to withdraw £100 and you were to process that
       withdrawal, it would then go into the stack and the net
       effect would change, wouldn't it, to be in green?
   A.  Yes.
   Q.  And that would be because there would then be cash due
       to the customer?
   A.  That's correct.
   Q.  Taking the two transactions together?
   A.  Yes, it would.
   Q.  Going back then to the failed transaction on 9 May, in
       your witness statement you seem to suggest in one
       paragraph -- and I will take you to it -- that once the
       £150 withdrawal was authorised, Horizon then instructed
       you to pay that amount over to the customer.
   A.  Yes, that's right.
   Q.  By that, when you use that kind of language, do you
       simply mean that the withdrawal went into the stack that
       we have just seen?
   A.  When the £150 withdrawal was done, the customer was
       inserting a card into the PIN pad on his side of the
       counter.  He would put the card in, when the transaction
       authorised it would bring "Authorised" on the PIN pad on
       the customer's side of the screen.  It would put the
       £150 onto the stack on my side of the screen and then it
       would print off a receipt off the Horizon printer and
       the receipt that came off there said "Authorised, £150"
       and then that's passed to the customer and then you pay
       the customer.  Press the "enter" button and pay the
   Q.  Sorry, press the "enter" key and then pay the customer?
   A.  Yes.
   Q.  That's the only point I wanted to clarify with you is
       that ordinarily what you would do -- I'm not talking
       about 9 May what you actually did, but what you should
       normally do is get all of the transactions, including
       any withdrawals, into the stack that we have looked at
       and then when the customer wants no more products or
       services you close the stack, or we would say submit it,
       by either pressing "enter" or "fast cash" or any other
       way that closes the basket and ends the session; that's
       right, isn't it?
   A.  If you do that you can't see what's actually -- you've
       got to then memorise then because as soon as you press
       that button and it goes there's no -- you've got no
       reference as what you're actually paying the customer
       because it will clear it, so you won't be able to see on
       the screen what you're supposed to be paying and if the
       customer then queries anything there will not be any
       information on your screen because it will have gone.
   Q.  I see, so what you say is you would prepare whatever
       money was due to the customer and then you may hand
       money over once you had the money in your hand, or you
       may click "enter" or "fast cash" and then hand it over,
       but the two are basically at the same time I think you
       are saying on a normal transaction?
   A.  You would normally take the money out, you would say it
       is £75 to pay or whatever -- like this case you say 150.
       You would have the 150, you count the 150, press the
       "enter" and it would clear.
   Q.  And on 9 May, the situation we're talking about, on the
       right-hand side of your screen here you would have had
       three transactions stacked up; that's right, isn't it?
   A.  Yes.
   Q.  You would have had the two for the customer earlier, the
       gentleman who was withdrawing money from two different
   A.  Yes.
   Q.  And you would have had the 150 from the customer who was
       presently before you?
   A.  Yes.
           Normally you wouldn't have two customers'
       transactions at the stack on one time.  It was just the
       circumstances on that day were exceptional.  The systems
       were slow, because they was connecting I had two
       customers' transactions on there, but I actually said
       that in my statement, that there was two customers.
   Q.  Yes, you have explained that you wouldn't ordinarily do
       it --
   A.  No.
   Q.  -- it wasn't the procedure you were supposed to
       follow --
   A.  Yes.
   Q.  -- but you were doing it in difficult circumstances?
   A.  That's right, that's correct.
   Q.  Could I ask you to turn to the witness statement of
       a Post Office witness, Ms van den Bogerd, which is at
       {E2/5/25}.  It should come up on in your screen.  Have
       you seen this witness statement before, Mrs Burke?
   A.  Yes, I have seen that.
   Q.  Can I ask you to read to yourself please paragraph 105.
       It largely confirms what we have just discussed.
   A.  Yes.
   Q.  Sorry, over the page as well please {E2/5/26}, just
       finishing paragraph 105.
   A.  Right, yes.
   Q.  Do you agree with what is said there?
   A.  I have read what she has said, but I have no idea of how
       the back office procedures of Horizon work.  We're
       trained on transactions and how to process transactions,
       we have no idea about how the system in the back
       connects to Horizon, when it goes to a bank, how it
       checks the bank or when it feeds things back.
   Q.  Sure.
   A.  We're not --
   Q.  Sure.  I'm not trying to suggest that you're aware of
       the technical data --
   A.  No.
   Q.  -- the data processing, everything that goes on within
       the system.  What I suggest you would have known is that
       it is when you close the stack that the transactions in
       the stack go into your accounts?
   A.  Yes, what should be in the stack, yes, should go into
       our accounts.
   Q.  And that happens when you close the stack, because
       that's when those transactions are finally confirmed?
   A.  When you press the "enter" button and it clears the
       screen, everything that's in the stack should be in your
   Q.  Thank you.
           On 9 May, you have explained that you were keeping
       a single session open and were processing withdrawals
       for two different customers from that single session
       because of the difficult circumstances.
   A.  Yes.
   Q.  You would ultimately though, wouldn't you, have to close
       that stack and, as we have just described, it is at that
       point that those transactions would go into the account?
   A.  Sorry, can you repeat that again please.
   Q.  You had kept the session open because you were concerned
       about problems with the system.
   A.  Right.
   Q.  But you would ultimately at some point, wouldn't you,
       have to close the stack in order to put the transactions
       finally through?
   A.  Yes.
   Q.  And so not closing the stack between customers but
       keeping the session open increased the time between you
       obtaining authorisation for the withdrawals and dealing
       with the customer and when the transactions then were
       submitted on the system; do you follow?
   A.  Yes, I don't think that would have made any difference,
       because it could have been the one customer doing those
       transactions himself anyway.  One customer could have
       been doing two card withdrawals and then ultimately
       a bank withdrawal, so that it was two customers doesn't
       really make any difference.
   Q.  It would have the advantage that the single customer
       would likely be there at the time you entered the
       transactions into Horizon; that would be one advantage
       of doing it the proper way?
   A.  No, because normally you have counted the cash out to
       your customer, pressed the "enter" key and then it is
       just on to the next customer -- the next customer comes
       forward then and you are serving then, so it would be
       the same whether it was one or two.
   Q.  I follow that, yes, Mrs Burke, but in relation to --
       thinking about the first customer, for whom the
       transactions actually were recovered, but thinking about
       that customer, when you ultimately come to press "enter"
       and clear the stack, it might be several minutes since
       those withdrawals had been authorised and since the
       customer had been in front of you; that's fair,
       isn't it?
   A.  But that would be the same if they were buying postal
       orders, stamps, et cetera, when -- you know, it could be
       the same time difference no matter what customer was
   Q.  Mrs Burke, to be clear, it's not the type of
       transaction.  The point I'm asking you about is the
       difference it makes that you were serving different
       customers from the same session so had to keep the
       session open, even after the time when the first
       customer may have left.  Do you follow that?
   A.  I don't see what you're trying to get at there.
   Q.  Well, imagine a really extreme circumstance where you
       kept the session open for an hour and served ten
       customers from it.  Do you follow?
   A.  Well, yes, but in a scenario you wouldn't do that.  You
       couldn't remember it.
   Q.  That's taking it to extremes, Mrs Burke, but the point
       is all I'm trying to suggest to you is that by not
       closing the stack as you go along after each customer
       session, what you're doing is increasing the gap in time
       between the authorisation and closing the stack and
       submitting the transactions.
   A.  You are increasing the time, but that could be the same
       if it had been the same customer who had carried on.
   Q.  I understand.
   A.  If it had been the first it would have been the same.
   Q.  Your point is if you had a customer who said he wanted
       to process a withdrawal and then had seven other
       transactions, he may have been in front of you for
       ten minutes.
   A.  Yes, it would have made no difference, it would still
       have been the same.  You would have still had that time
       delay before you actually cleared your stack and then
       they would have gone onto your log.
   Q.  In that circumstance though it's right, isn't it, that
       you would not, following proper procedure, have passed
       the money to the customer before you had gone through
       the rest of the transactions?
   A.  You would complete all the transactions that that
       customer was wanting before you cleared.
   Q.  Yes.  And on 9 May it was when you went to settle the
       basket and clear the screen that the system went down,
       wasn't it?
   A.  Yes.
   Q.  And Horizon then printed a disconnected session receipt
       and we can see that at {F/1461} please.  What that
       receipt shows is essentially what was in your stack in
       that session when the system went down; that's right,
       isn't it?
   A.  That's correct, yes, but that took two minutes to come
       up to after I had been serving the customer.
   Q.  Yes.  And then what you next saw -- I'm just running you
       through the documents.  What you next saw is the
       document at {F/1464}.  This again, Mrs Burke, was a few
       minutes later this receipt was printed; that's right,
       isn't it?
   A.  Correct.
   Q.  And you say in your witness statement that what you
       initially understood from this receipt was that the two
       transactions listed on it had failed to recover; that's
       right, isn't it?
   A.  Yes.
   Q.  And you inferred from that that the third transaction
       for £150 had been successfully recovered?
   A.  Yes.
   Q.  You now understand, it seems from your witness
       statement, that what this receipt was intended to convey
       was two things.  1, there had been a failure of recovery
       for some transactions, one or more, but that these two
       listed had been recovered?
   A.  Yes.  I found that out when I went through my
       transaction log and I started to tick off the
       transactions, then I realised that the "recovery failed"
       wasn't the fail -- the items it showed weren't the
       failed transactions, they were the ones that had
       actually gone through and it was the other transaction
       that had failed.
   Q.  Yes.  Shall we look at the -- the transaction log is at
       {F/1465}.  So you printed this off within a matter of
       minutes, as I understand it, is that right?
   A.  It was after I had closed the branch.
   Q.  So approximately how long, do you think?
   A.  I did say in my statement.  Can I just refer to it?
   Q.  Well, the "recovery failed" receipt was printed at 9.36.
       We see that at paragraph 17 of your statement {E1/4/3}.
       You don't actually say at paragraph 17 --
   A.  9.36, yes.
   MR JUSTICE FRASER:  You phoned the helpline after you had
       closed the branch?
   A.  Yes.  I served another couple of customers and there
       were some problems with connection then and then
       I decided to close the branch and then I went and ran
       off the transaction log and that's when I sat down,
       closed the branch and worked through it.
   MR JUSTICE FRASER:  And at the end of paragraph 18, did you
       then phone the helpline?
   A.  I phoned the helpline -- I think it was --
   MR JUSTICE FRASER:  So we can work the time window out,
       I think.
   MR DRAPER:  Yes, my Lord.
           Looking at this document that you have on screen at
       the moment, Mrs Burke --
   A.  It would have been about 9.34.  It would have been after
       the -- because that's -- when the 253 total came up ...
   Q.  In any event, the precise time is --
   A.  About -- say 9.34.
   Q.  Yes, so you printed off this transaction log that we can
       see on screen and we can see from your handwritten
       annotation on it the ticks on the transactions that you
       expected to see and do see there and you have then
       indicated between two transactions that the £150
       withdrawal was missing.
   A.  Yes.
   Q.  So you were able to work out quite quickly from looking
       at this transaction log that one of the transactions you
       expected to see wasn't there?
   A.  Yes.
   Q.  It's from that that you corrected what you had
       understood from the "recovery failed" receipt?
   A.  Yes.
   Q.  You saw it was the other way round?
   A.  It was the other way round, yes.
   Q.  So by this point, after you have sat down with the
       transaction log, you knew you had had a system outage,
       problem with the system, you knew that the session you
       were in when the outage occurred had a problem, because
       you had the discontinued session receipt?
   A.  Yes.
   Q.  You knew that the recovery process on Horizon, the
       automatic recovery process, had failed in some way; you
       knew that from your "recovery failed" receipt?
   A.  Yes.
   Q.  And you had been able to identify from this transaction
       log which transaction it was that had not been
   A.  Yes.
   Q.  And that meant that in this fairly short space of time
       you were able to work out that the £150 withdrawal for
       which you had physically paid out had not gone into your
   A.  Yes.
   Q.  And you knew that the result of that was that unless you
       obtained a transaction correction to put the transaction
       into your accounts, you were going to have a £150
   A.  Yes.
   Q.  And you knew what would have caused that shortfall,
       didn't you, you knew it was the failed recovery for the
   A.  Yes.
   Q.  Can I ask you to look at paragraph 19 of your witness
       statement please {E1/4/4}, and can I ask you to look in
       particular, Mrs Burke, at the last sentence of that
           Have you read that?
   A.  I'm just finishing it off.
   Q.  Okay.
   A.  Right.
   Q.  What I suggest to you, Mrs Burke, is that that sentence
       isn't accurate, is it?  You had been able to ascertain
       from the system what the discrepancy would be, the £150
       shortfall, and you knew exactly what its cause was?
   A.  But it hadn't been recorded on my transaction log, so it
       had not shown that a transaction for that amount had
       gone through on my accounts.
   Q.  Yes that's right, but that actually was something that
       helped you understand what had happened, wasn't it?
       That's what showed you that it hadn't got through to
       your accounts.  Do you follow?
   A.  I just knew that the system -- it was missing off my log
       and that it should have been there and then my accounts
       would have balanced, but as it was it wasn't there, so
       my accounts didn't balance.
   Q.  Yes, I appreciate that.  I'm focusing quite closely, if
       I may, on the wording of the last sentence of
       paragraph 19 of your statement.  It says there:
           "There was therefore no means through the Horizon
       system for the discrepancy to be identified ..."
           Stopping there, I suggest to you that's wrong
       because you knew exactly what the discrepancy would be,
       it would be a £150 shortfall.  So that's not correct, is
   MR JUSTICE FRASER:  Well, Mr Draper -- well, I will wait for
       the answer and then ...
   A.  Well, the Horizon system didn't identify it; I knew it
       was the £150.  Oh, I see what you mean, that I had the
       two receipts.  Yes, I will agree with that.
   MR DRAPER:  And focusing on the second half please of the
           "... or for its cause to be established in [your]
           I suggest that based on what we have just discussed,
       you knew what the cause of a £150 shortfall would be,
       didn't you?
   A.  Sorry, could you say that again?
   Q.  You knew that when a £150 shortfall arose --
   A.  Yes.
   Q.  You knew what would have caused it, because you had
       worked out that the £150 withdrawal hadn't made its way
       to your account?
   A.  I did after studying all that paperwork, yes.
   MR JUSTICE FRASER:  Mr Draper, just so I understand, by the
       discrepancy being shown on the Horizon system you mean
       the receipt, do you, that was printed off, or do you
       mean something else?
   MR DRAPER:  No, the point I was taking was "no means for the
       discrepancy to be identified" and my point was that
       Mrs Burke had identified the discrepancy.
   MR JUSTICE FRASER:  Well, you actually were putting the
       point on the last sentence which says "no means through
       the Horizon system for the discrepancy to be
   MR DRAPER:  Yes, that's right.
   MR JUSTICE FRASER:  So are you putting the point that it was
       the receipt that demonstrated the discrepancy being
       identified in the Horizon system, or is it in some other
   MR DRAPER:  Well, two points, if I may.
           First, Mrs Burke, you knew there would be
       a shortfall without even asking Horizon to tell you
       that, didn't you?
   A.  Because I had paid the customer the money.
   Q.  Yes.  So that's the first point.
   A.  And it wasn't recorded on my transaction log.
   Q.  And secondly, you could see that discrepancy, couldn't
       you, by running a balance snapshot?
   A.  Because the transaction was missing on the snapshot.
   Q.  Yes.  And if you were -- we will come back to it,
       my Lord.
           The next step then, after what we have just gone
       through, is that you call the helpline; that's right,
       isn't it?
   A.  That's right.
   Q.  You can see the transcript at {F/1466}.  Before we look
       at the detail of it, it's this call to the helpline that
       makes you think that without your efforts in tracking
       the customer and taking him to the bank and so on,
       Post Office would not have resolved the problem for you,
       is that fair?
   A.  That's fair, yes.
   Q.  Mrs Burke, I assume you are quite familiar with the
       transcript.  Do you need an opportunity to read it, or
       would you be happy for me to ask you questions about it?
   A.  I think I would be happy.  There will be odd bits.
       I can always refer back to it if I'm not.
   Q.  You are the customer shown on this transcript, aren't
       you?  Mrs Burke?
   A.  Sorry?  Oh, yes, I'm the customer.
   Q.  And if we look to the bottom of this first page -- this
       is you speaking -- you explain in that final entry on
       the page that you know the £150 withdrawal hasn't gone
       through and you know that your branch therefore won't
   A.  Yes.
   Q.  If we flip over the page now please {F/1466/2}, you then
       explain there about the disconnected session.  I think
       it is fair to say, just skipping forward, you didn't
       feel you were getting very far with the initial helpline
       operator so you asked to speak to a supervisor?
   A.  Correct, yes.
   Q.  If we go then to the conversation with the supervisor
       and pick it up on the third page please {F/1466/3}, do
       you see four entries from the bottom there's an entry
       starting "But the £150"; do you see that?
   A.  Yes.
   Q.  The supervisor says there:
           "But the £150 hasn't gone through, that's why it's
       not on your transaction log.  If the £150 hasn't gone
       through and you have given the customer the £150 that
       will give you a discrepancy because there is no
       transaction for it."
           So that reflected your understanding, didn't it?
   A.  That's right.  But she is saying there's no transaction
       for it, but it had already given me an authorised
       receipt and the customer had had an "Authorised" message
       on the screen as well, on the PIN pad.
   Q.  Yes, we will be I think just several more minutes on
       this transcript.
           On page 4 --
   MR JUSTICE FRASER:  Sorry, do you mean you want to keep
       going past 1 o'clock or ...
   MR DRAPER:  Yes just the next three minutes.
   MR JUSTICE FRASER:  For how long?
   MR DRAPER:  Three minutes, my Lord.
   MR JUSTICE FRASER:  Yes, of course.
   MR DRAPER:  On page 4 though please {F/1466/4} in the middle
       do you see a long paragraph where the supervisor starts
       by saying "No"?
   A.  Yes.
   Q.  The supervisor there says:
           "When you're doing the transaction, at the very end
       of the transaction.  When you're selecting 'settle'
       before you give the customer any money, because you
       should never give the customer any money until the
       transaction has been settled, it's at that point that
       the transactions seem to be disconnecting.  It's
       a nationwide issue.  Erm, we are aware of it and it's
       being dealt with by the Horizon service desk at the
       moment.  If the transaction isn't on your transaction
       log then the transaction hasn't gone through."
           We see that you're not very pleased with that
       answer.  And then the supervisor says:
           "The only thing we can advise at this point, if you
       want to discuss this further you would need to speak to
       the IT help desk ... because they are the ones that are
       running the invest ..."
           And I suggest that would be "investigation".  Does
       that accord with your recollection?
   A.  Yes.
   Q.  Then towards the bottom of the page you see the third
           "Yeah but it is a nationwide issue at the moment and
       we are aware of it.  So it may be -- I can't guarantee
       anything but it may be once we are aware of it and we
       can sort something out.  It may be we are looking into
       it at a later date to try and recoup any losses but we
       can't guarantee anything at the moment until we are
       aware of what is causing the issue."
           So two points I would like to take from what we see
       from this transcript, Mrs Burke.  The first is that you
       had been informed it was a nationwide problem that was
       being investigated?
   A.  Yes, it was a nationwide problem.
   Q.  And that it was being investigated?
   A.  But they were also saying that they can't guarantee
       anything either.
   Q.  Yes.  But you accept that you were told it was being
       investigated and you understood that?
   A.  I did, but never at any point did they say in the
       conversation that they would come back to me and take
       a note of my problem and "We will come back to you and
       make a note this is the problem you're having in your
       branch", which --
   Q.  Maybe I will try and summarise the position.
   A.  Okay.
   Q.  I appreciate that this call didn't give you the
       certainty that you would have liked that you were going
       to be taken care of and the problem resolved
   A.  Yes.
   Q.  But it would be unfair, wouldn't it, to suggest that
       Post Office told you that it wasn't going to resolve the
       situation.  It essentially said that that might be the
       case.  Is that fair?
   A.  They are saying they couldn't guarantee and for me as
       a branch, if they can't guarantee, I would be losing --
       I would be out-of-pocket.
   Q.  Yes and you understood from what you had been told here
       that there were people other than the helpline operators
       here, you were told about the IT service desk,
       essentially this was being investigated by someone else
       and they would get back to you with the results; that's
       fair, isn't it?
   A.  They didn't say they would get back to me.  I can't
       recall them saying they would come back to me.
   Q.  If you look at the third entry from the bottom,
       Mrs Burke -- it is not precisely in those words -- third
       entry from the bottom, last sentence:
           "It may be we are looking into it at a later date to
       try and recoup any losses ..."
           That would be recoup losses for you, wouldn't it?
   A.  They are saying that they may be looking into it.
       That's not a guarantee of looking into it.
   Q.  I understand.
           I think we can sense apply take a break there,
       my Lord.
   MR JUSTICE FRASER:  Right, we're going to have a break for
       lunch.  I would like you to come back at 2 o'clock
       please.  Because -- I say this to every witness so don't
       read anything into it, but because you are in the middle
       of giving your evidence don't talk to anyone about the
       case please and come back at 2 o'clock.
           Anything else?  No.  2 o'clock.  Thank you all very
   (1.05 pm)
                    (The luncheon adjournment)
   (2.00 pm)
   MR DRAPER:  Mrs Burke, we have just got through before the
       break the helpline call, going through the transcript.
       The next step that you explain in your witness statement
       is the things that you yourself did to track down the
       customer and obtain records from him and from his bank.
   A.  That's correct, yes.
   Q.  And it's right, isn't it, that it was on 12 May that you
       went with the customer to his bank to obtain
       confirmation of the withdrawal?
   A.  That's right.
   Q.  And it is one day after that, on 13 May, that your
       husband spoke to a lady at Post Office and he was told
       in that telephone call that your branch could expect to
       receive a transaction correction?
   A.  Yes.
   Q.  Presumably it was in that call, on the 13th, that you or
       your husband told Post Office about everything that you
       had done in the meantime?
   A.  Yes, we told them that we had seen -- that we had seen
       the "authorised" receipt from the customer and that we
       had also got the statement from the bank to show that
       the money had gone out of the customer's bank account,
       even though it had not appeared on our system.
   Q.  I should make clear before moving on that Post Office
       makes absolutely no criticism of you for doing all that
       work yourself and for having tracked the customer down
       and similarly I didn't intend to criticise you earlier
       about the impression you took from the telephone call.
       You were understandably frustrated.  I just wanted to
       make that clear.
   A.  Fine.
   Q.  The point I do intend to challenge in your witness
       statement is a relatively small one, which is your
       suggestion that the shortfall in your branch would not
       have been resolved if you hadn't done the work that you
       did.  Do you follow?
   A.  I understand what you're saying.
   Q.  It might help to take it from your statement at
       paragraph 27, which is {E1/4/5}.  It is the second
       sentence in that paragraph, Mrs Burke.  That's the only
       one I want to address with you at the moment.  You say
           "Based upon my initial experience of the helpline,
       I do not think that Post Office would have resolved this
       if I had not had the clear proof that the £150
       transaction had in fact been authorised and that the
       money had left the customer's bank account."
   A.  No.
   Q.  You of course were not aware at the time in May 2016, or
       indeed when you prepared this witness statement, of the
       steps that were going on in the background at
       Post Office and at Fujitsu, were you?
   A.  No.
   Q.  What I intend to do is show you some documents produced
       during that process that I will suggest show that
       Post Office and Fujitsu were dealing with the problem
       and would have been able to resolve it.  Do you follow?
   A.  I follow.
   Q.  The first document is at {F/1470.1}.  This is what is
       called a BIMS incident report which is a wonderfully
       tautologous phrase because BIMS stands for business
       incident management service, which is a function
       performed by Fujitsu.
           You see at the bottom that this is a document
       produced by Mr Andy Dunks.  Do you see that in the
       rectangle at the bottom of the page "Analyst" on the
   A.  Yes.
   Q.  You can take it from me that he is a Fujitsu employee.
       This is a report that Fujitsu has generated to identify
       branches affected by the outage on 9 May.  Again, in the
       box at the bottom of the page do you see under "actions:
       date and time", the first entry there:
           "On the 9th May 2016 there was a major outage at
       branches that caused multiple transactions to fail.
       Transactions that require [Post Office Limited]
       reconciliation have been identified and a spreadsheet
           Do you see that?
   A.  Yes.
   Q.  And do you see that this document is dated 12 May,
       that's in the same box after "Actions: date and time"?
   A.  Yes.
   Q.  So am I right to say this is then the same date as that
       on which you went to the customer's bank with him?
   A.  Yes.
   Q.  Going right to the bottom of this document, do you see
       at the bottom it says:
           "The resultant spreadsheet has been attached to the
       same email that this BIMS has been sent."
   A.  Yes.
   Q.  I can then show you that email quickly.  It is at
       {F/1470.1.1}.  So this is the email that attaches the
       report that you have just seen, Mrs Burke, and you see
       from the top of the email that it is sent to various
       recipients at Post Office.
   A.  Yes.
   Q.  And it says in the last sentence of the email:
           "I have attached a spreadsheet containing the list
       of all transactions that require POL ..."
           Which is Post Office Limited:
           "... to manually reconcile as appropriate."
   A.  Yes.
   Q.  The next document, Mrs Burke, is a spreadsheet which is
       at {F/1848.4}.  It will just take a moment.  This,
       Mrs Burke, is the spreadsheet that we have just seen
       described as having been attached to the email that went
       to Post Office.  If you could please look at row 12, if
       the operator could scan down please.  You recognise in
       the first column, column A, where it identifies the
       branch, the branch ID there 216321.  Do you recognise
       that branch ID?
   A.  Yes.
   Q.  That's the Newport branch, isn't it?
   A.  The Newport branch, yes.
   Q.  Now can we move it over just slightly so we can see all
       of column H please.  What it says there in this Fujitsu
       document is:
           "The £150 cash withdrawal transaction was authorised
       by the FI ..."
           Which I take to be financial institution:
           "... and an authorised receipt was produced on the
       counter.  However, when the user attempted to settle the
       transaction it failed due to the known data centre issue
       at the time so disconnected session receipts were
       produced and the user was logged off.  The user managed
       to log back in but recovery also failed.  As
       an authorised receipt was produced the user should have
       handed money over to the customer but we cannot be
       certain that they actually did so.  Assuming money was
       handed over the customer account will be correct but the
       branch will have a shortage given that the transaction
       hasn't been recorded on the system.  This will need to
       be manually reconciled."
           So what I suggest is it looks, doesn't it, like
       whoever at Fujitsu wrote this didn't know whether or not
       you had paid money over the counter, is that right?
   A.  Yes, they have put they don't know if it was handed over
       even though I had already called in and said we had paid
       the customer.  So they knew on the 9th that I had paid
       the customer.
   Q.  Remember, Mrs Burke, this is a Fujitsu document rather
       than a Post Office document so this is the people
       managing the system and what I suggest is this is what
       they have discerned from the data available to them on
       the system.  Do you follow?
   A.  Right.
   Q.  And that they have essentially obtained data reports on
       the thousands of branches that were affected and this is
       what they can see from their data.
           So do you see then that Fujitsu says, on the
       assumption that the money was handed over,
       reconciliation would be required?
   A.  Right.  Yes, I can see what they're saying.
   Q.  Continuing with this account of what's going on in the
       background, Post Office then produces a spreadsheet to
       show what it is doing in response to this analysis from
       Fujitsu and that's at {F/1848.5}.  It is another
       document that will take a moment to come up.  Do you see
       there, Mrs Burke, row 13 that's been highlighted?
   A.  Yes.
   Q.  And we see in column A the branch code but helpfully the
       branch name in column B as well.  Then if the screen can
       be moved over a little please.  Column H is the same, it
       is the statement we have already seen from Fujitsu that
       we just read.  If we move further please to the right --
   A.  When it says the transaction date and it says 9 May.
   Q.  Forgive me, where are you looking?  Can we move the
       screen back.
   A.  9 May.  But it says the time 8.28.  How can that be
       8.28?  My branch doesn't open until 9.
   Q.  I'm instructed that these records will be an hour
       different because of not accounting for savings time.
       So it's just every time you see a document like this
       they will be an hour out.
   A.  Earlier?
   Q.  Forgive me?
   A.  An hour earlier?  So how would they know I had done that
       an hour earlier?
   Q.  No, the suggestion isn't that the transaction occurred
       at 8.28.  It did occur -- the session start date I think
       that will show.  No one is suggesting that it was at
       8.28, it's just this time stamp doesn't take account of
       changes in GMT, so it will be an hour off.  This time
       stamp is an hour wrong.
   MR JUSTICE FRASER:  I don't think GMT does change.
   MR DRAPER:  Forgive me.
   MR JUSTICE FRASER:  I think what you mean is this is GMT and
       it would be daylight saving time in May.
   MR DRAPER:  That's right.  I'm sure that's what I meant.
   MR JUSTICE FRASER:  So it has GMT at the top of the
       column under letter F.
   MR DRAPER:  Yes, I think it is just unadjusted.
   A.  Okay.
   Q.  If we can move across then please.  I think you
       confirmed that H looks like the same content we have
       already seen, Mrs Burke.  If we go back just across
   A.  Yes.
   Q.  That's what we have just already seen in the other
   A.  Okay.
   Q.  But this is a document into which Post Office has added
       further information and that's over in column J.  You
       see the heading to that column is "Action taken" and
       what Post Office has included in that box for your
       branch is:
           "Branch contacted.  Shortage in branch so TC issued
       to adjust cash."
           So what I suggest is that's Post Office, with the
       information it has about what actually happened in your
       branch, information from you, confirming that it will
       issue a TC?
   A.  Right.
   Q.  And I suggest that this coincides, Mrs Burke, with your
       witness statement where you say that your husband was
       told on 13 May to expect a transaction correction.
   A.  Yes.
   Q.  Just to complete the picture, the transaction correction
       was in fact produced on 16 May.  We can see that from
       {F/1687.1}.  If you look please at row 5 you will see
       the date at A, 16/05/2016.  And then if the operator
       could move the screen along a little please to the
       right, in column L we see a description -- we see the
       text that accompanied the transaction correction and
       that reads:
           "To correct communications failure on 9.5.16 for in
       pounds 150 ..."
           It gives a reference and it says then "so credit to
       office", then provides the number for the financial
       services enquiry team.  That's the transaction that you
       received -- I think you say received and processed on
       the 17th, is that right?
   A.  Yes, 17 May.  Yes, it was 17 May when we got it through.
   Q.  So the point I want to take from these documents that
       I have just taken you through, Mrs Burke, is nothing to
       do really with the impression that you took at the time
       or the impression that you set out in your witness
       statement -- as I say, we don't criticise you for having
       taken that view -- the point I'm trying to establish is
       that having seen now, with the benefit of what you have
       seen here about what Post Office and Fujitsu were in
       fact doing in the background, as it were, unknown to
       you, with the benefit of that knowledge would you accept
       that it looks as though Post Office would have been able
       to resolve the problem in your branch even if you hadn't
       taken the steps that you did?
   A.  Possibly, yes.
   Q.  Thank you.
           No further questions, my Lord.
                    Re-examination by MR GREEN
   MR GREEN:  Only one question, my Lord.
           Mrs Burke, if you just have a very quick look on
       page {F/1461/1}.  Do you recognise that?
   A.  Yes.
   Q.  The transactions that we see on there, are they in
       chronological order going down, so 180 comes first,
       73 --
   A.  Then 73, then 150.
   Q.  -- or are they going up in chronological order?
   A.  No, the 180 and then the 73 and the 150 was the last
   Q.  Right.  And that's the one there was a problem with.
   A.  Yes.
   Q.  Thank you very much.
   MR JUSTICE FRASER:  Thank you very much, no questions from
       me.  You are free to go, thank you.
   MR GREEN:  My Lord, there is a tiny possibility I might need
       to ask your Lordship for just five minutes.  There is
       a PEAK that has been uploaded, a new one today, which
       I don't understand and the haste with which it has been
       done suggests it might be going to be put to Mr Roll.
   MR JUSTICE FRASER:  When you say uploaded you mean put in
       the trial bundle?
   MR GREEN:  Yes.  And so because I haven't had a chance to
       discuss it with him, can I have five minutes?
   MR JUSTICE FRASER:  When did that appear, that document?
   MR GREEN:  During the course of today.  It uploaded at 9.43
       this morning but unfortunately I haven't managed to get
       on top of it.
   MR DE GARR ROBINSON:  My Lord, I won't be asking any
       questions on that PEAK.
   MR JUSTICE FRASER:  You're not?  You are cross-examining
       Mr Roll, are you?
   MR DE GARR ROBINSON:  I am cross-examining Mr Roll and there
       will be no questions about that PEAK.  It is there for
       an entirely separate, coincidental reason.
   MR JUSTICE FRASER:  Is Mr Roll going into tomorrow?
   MR DE GARR ROBINSON:  Yes, I would expect so.
   MR JUSTICE FRASER:  Are there going to be any questions on
       that PEAK tomorrow?
   MR JUSTICE FRASER:  No, all right.
           Well then that means there won't be a situation
       where you need to speak to Mr Roll overnight and he will
       be in purdah.
   MR GREEN:  My only anxiety, my Lord, is if it arises in the
       evidence of someone else and I haven't dealt with it --
       I don't want to be -- could I just have five minutes to
       make sure?
   MR JUSTICE FRASER:  I'm not going to give you five minutes,
       I will actually give you ten minutes.  That's the
       easiest way of dealing with it.  So we will come back in
       at half past.
   (2.20 pm)
                          (Short Break)
   (2.30 pm)
   MR GREEN:  My Lord, I call Mr Roll if I may please.
                    MR RICHARD ROLL (affirmed)
   MR JUSTICE FRASER:  Thank you, Mr Roll.  Have a seat.
   A.  Thank you.
                 Examination-in-chief by MR GREEN
   MR GREEN:  Mr Roll, in front of you there should be
       a folder, you can see.
   A.  Yes.
   Q.  And if you turn please to tab 7 {E1/7/1}.
   A.  Yes.
   Q.  You can see there's a witness statement with your name
       on it.
   A.  Yes.
   Q.  And at the back of that witness statement is
       a signature.
   A.  Yes.
   Q.  Is that your signature?
   A.  It is.
   Q.  And are the contents of that witness statement true?
   A.  Yes, they are.
   Q.  And if we go to tab 10 {E1/10/1}, have you got a copy of
       the amended statement there?
   A.  Not here, no.  Thank you.
   Q.  Do you want to keep it in front of you.
   A.  Thank you, yes.
   Q.  Have a look at the back page of that if you would, very
   A.  Yes.
   Q.  And have you got your signature?
   A.  Yes.
   Q.  On that one, or have you signed another one?
   A.  I have signed another one, yes.
   Q.  And are the contents of that statement true?
   A.  Yes.
   Q.  Thank you very much.
             Cross-examination by MR DE GARR ROBINSON
   MR DE GARR ROBINSON:  Mr Roll, you say that between 2001 and
       2004 you worked at Fujitsu and you describe it in
       paragraph 2 of your first statement {E1/7/1} as working
       in third/fourth line support.
           I would just like to ask you some general questions
       about the support teams first of all and for that
       purpose I would like to ask you to look at Mr Parker's
       first witness statement, the reference is {E2/11/5}.  It
       will come up on the screen for you.  You will see that
       Mr Parker says in paragraph 24:
           "There were four lines of support for Horizon while
       Mr Roll was employed by Fujitsu and they are described
       in paragraph 26 below.  There are still four lines of
       support ..."
           Stopping there actually there is a general question
       I should ask which is that you know that Mr Parker has
       made three witness statements, don't you, in this
   A.  I know he has made some witness statements, I don't know
       how many.
   Q.  Have you read any of them?
   A.  I have looked at them briefly.
   Q.  Have you looked at all of them briefly or just some of
       them briefly?
   A.  I don't think I have read all of them.
   Q.  Okay.  His first witness statement, the witness
       statement that we're now looking at, have you read that
       carefully or briefly?
   A.  Briefly.
   Q.  Paragraph 25, perhaps I could ask you to read that
       paragraph very quickly to yourself.
   A.  Yes.
   Q.  That's true, isn't it, what's said there?
   A.  Yes.
   Q.  And then he describes in paragraph 26 the four lines of
       support for Horizon.  The first line involved several
       different elements.  This is in paragraph 26.1.  First
       of all he refers to the Horizon service desk, which was
       a help desk operated by Fujitsu that branches could
       contact with issues relating to Horizon application or
       hardware provided in branch.  Then over the page
       {E2/11/6}, in 26.1.2 he said:
           "There was also a 1st line communications management
       team operated by Fujitsu which specifically focused on
       communication incidents ..."
           And then finally in 26.1.3 he says:
           "Post Office also operated a 1st line help desk for
       operational issues called the national business support
       centre ..."
           Are you able to confirm that all those facts are
       true from your own knowledge?
   A.  I don't remember.
   Q.  You do know that Post Office operated a first line help
       desk for operational issues called the national business
       support centre, don't you?
   A.  There was a first line support desk and I believe it was
       shared between Fujitsu and the Post Office, or ICL and
       the Post Office as it was when I joined, but I don't
       recall what the splits were or anything like that.
   Q.  That's fair.
           Then if I could ask you to read the next unnumbered
       paragraph underneath 26.1.3 to yourself please.
   A.  Yes.
   Q.  Are you able to agree with the points made in that
   A.  From what I remember, yes.
   Q.  I'm grateful.
           Then the next unnumbered paragraph, Mr Parker says:
           "If NBSC were unable to identify the cause of
       a discrepancy they would often fall back on a default
       statement along the lines of 'this looks like a software
       issue' so that the SSC would investigate it.  However,
       Mr Roll's statement that 'if an error was referred to us
       then it was extremely unlikely to be due to a mistake
       made by a postmaster' is not correct.  The vast majority
       of discrepancies investigated by the SSC as pseudo
       'software issues' were (and are) not caused by software
           I would like to suggest to you, Mr Roll, that that's
       true, isn't it?
   A.  The way that I remember it, it was issues to do with the
       software that were causing the problems.  Whether that
       was the programme that had been written or data
       corruption, that's what I remember as our problems
   Q.  You remember there being problems with data corruption?
   A.  Yes.
   Q.  I'm not asking you about data corruption, Mr Roll, I'm
       asking you about software issues.  The claim I would
       like to put to you again is the last sentence of that
           "The vast majority of discrepancies investigated by
       the SSC as pseudo 'software issues' were (and are) not
       caused by software issues."
           Are you in a position to agree to that?
   A.  From my recollection I would disagree with that, but it
       was a long time ago.
   Q.  I'm grateful.
           Paragraph 26.2 {E2/11/7}, there's then a description
       of the second line support:
           "Second line support was provided by senior members
       of the HSD and SMC ..."
   MR JUSTICE FRASER:  Hold on, the page hadn't changed.
   MR DE GARR ROBINSON:  I'm so sorry, it is page 7.
   MR JUSTICE FRASER:  It is all right.  It has now, it just
       takes a little while.  We are at 26.2, yes?  Or
       whichever one it is on page 7.
   MR DE GARR ROBINSON:  Can I ask you to read paragraph 26.2
       and its description of the second line of support.
   A.  Yes.
   Q.  Is that something that you can recall and can agree
   A.  It's not something I can definitely say yes or no to.
   Q.  Your caution does you great credit, Mr Roll.  What
       I discern from you -- and if I'm putting words in your
       mouth do let me know -- you think it may well be right
       but you don't want to commit yourself --
   A.  Yes.
   Q.  -- because it is a long time ago and you didn't work in
       second line?
   A.  Yes.
   Q.  That's very fair.  But just to be clear, people who
       worked in the second line included junior members of the
       SSC team, yes?
   A.  Hm.  That's not how I remember it.  Second line was
       a separate entity and we were third line in the SSC.
   Q.  Well, I put his evidence to you and I will move on.
           Then we come to the third line, 26.3.  He says:
           "The SSC also provided 3rd line support.  The staff
       that provided 3rd line support had a detailed knowledge
       of the Horizon application based on documentation and
       some inspection of source code.  They:
           "Designed, tested and documented work rounds for the
       1st and 2nd lines of support;
           "[They] applied analytical skills to the symptoms
       and evidence gathered by the 1st and 2nd line functions
       and undertook in-depth investigation into incidents
       (incidents are the basic unit of work for the support
       team and come from help desk calls and other Horizon
       support teams);
           "[They] undertook complex configuration
       (configuration items can be used to alter the behaviour
       of the application) and data fixes which might have
       required the generation of special tooling;
           "[They] designed, wrote and documented new support
       tools ... [they] undertook source code examination,
       complex diagnosis and documentation (including methods
       to recreate faults) of new application problems before
       sending them to the 4th line support group for root
       cause software fix; and [they] provided technical
       support to other internal Fujitsu teams working on
           Now, I have read an awful lot to you, Mr Roll, and
       I have done that on the apprehension that none of this
       is going to be controversial.  Can I ask you to indicate
       whether you accept what Mr Parker says there in
       paragraph 26.3?
   A.  Yes.
   Q.  I'm grateful.  Then there is a description of the PEAK
       and KEL system which I need not trouble you with.
       Paragraph 26.4 on page 8 {E2/11/8}, there's
       a description of the fourth line:
           "4th line support staff had an intimate knowledge of
       narrow areas of the system and were (and are) ultimately
       responsible for the production of permanent fixes to
       repair the root cause of an incident or problem in the
       live application.  They had knowledge of computer
       languages which they used to amend source code to fix
       problem in the live application code.  There was often
       overlap between 4th line and developers, who added new
       features into the application."
           Do I apprehend, Mr Roll, that you agree with that as
   A.  Broadly speaking, yes.
   Q.  I'm grateful.
           Now, you describe yourself in paragraph 2 of your
       first statement {E1/7/1} as "third/fourth line support".
       You in fact worked in third line support, didn't you?
   A.  I worked in third line support, yes.
   Q.  And it is fair to say that -- well, this is my language
       and if you disagree with the nomenclature do let me
       know, but would it be fair to say that third line
       support was the elite, but the fourth line support, the
       people that actually fixed software problems, they were
       the super elite.  Would that be a fair thing to say?
   A.  I would say there were some members of third line who
       were super elite as well.
   Q.  And would you describe yourself as one of those people?
   A.  No.
   Q.  Let's talk about third line.  The third line, they were
       on a floor that even other Fujitsu staff members
       couldn't generally access, is that right?
   A.  Yes.
   Q.  And if we can pick it up at paragraph 28 of Mr Parker's
       witness statement {E2/11/8}, you worked in third line
       for between -- well:
           "Between 1 January 2001 and 31 December 2014 ..."
           And those are the years that you worked for Fujitsu:
           "... the SSC received a total of 27,005 calls,
       meaning that on average 563 calls per month were dealt
       with over this 4-year period."
           And he refers to a spreadsheet setting that out and
       he then analyses the data in that spreadsheet.  Would
       you accept that that is a fair reflection of the amount
       of calls coming in, the amount of incidents coming in to
       SSC when you were there, third line?
   A.  I can't really remember.  I know there were periods when
       it was very busy and periods when it wasn't so busy.
       Sometimes we had three or four jobs on the go at once,
       other times we were given other work to do from the
   Q.  And paragraph 29, Mr Parker says:
           "Transferred calls (ie those not resolved by the
       SSC) are of interest."
           He says:
           "A very small proportion of calls transferred to
       4th line support would have concerned software errors
       requiring resolution ..."
           Stopping there, Mr Roll, that's true, isn't it?
   A.  Yes.
   Q.  So he then says:
           "... it would be interesting [therefore] to know the
       number of calls transferred to [fourth line]."
           Would you agree to that?  It would give some
       indication of the extent to which incidents coming into
       the SSC properly, genuinely represented software areas
       that required fixing?  That would be a useful way of --
       a touchstone of trying to work out --
   A.  Yes.
   Q.  Thank you.  He then says that, unfortunately:
           "... while the SSC have records of the volume of
       transferred calls, we do not retain records of where
       they are transferred to and it is not the case that all
       of these would have been transferred to 4th line
       support.  For example incidents would often arrive at
       SSC from internal teams for routing back to help desks."
           Do you remember that?
   A.  I don't remember that particular ...
   Q.  But would it be right to say that of the calls coming
       into third line support, a significant proportion would
       go -- of calls that would then be transferred out, would
       go to places other than fourth line support?  Would that
       be fair?  Does that accord with your recollection?
   A.  The way I recollect it is that calls would come in and
       we would work on them, either fix them, in which case
       they would go back to the originator, or we would pass
       them on.
   Q.  To other people?
   A.  To other people.
   Q.  Depending on the nature of the problem?
   A.  Yes.
   Q.  So if there was an infrastructure problem you would
       probably pass the call on to the infrastructure team, it
       wouldn't go to fourth line support, would it?
   A.  Probably.
   Q.  And there are a number of other teams that calls coming
       into third line would be passed to who would not be
       fourth line support?
   A.  Probably.  I can't remember the full details.
   Q.  I accept, it's a long time ago.  If we could just look
       at an analysis that Mr Parker has done of calls coming
       in and calls coming out, could we go to {F/1839} please.
       This is a series of spreadsheets which I have -- I'm
       probably the least qualified person in court to be
       talking about spreadsheets and I would like to go to the
       tab which is "Live PEAKs into SSC" please, at the bottom
       of the page.  "RRP live PEAKs into SSC", there.  And you
       will see that there is a calculation of from 2002 -- oh,
       from 2002 all the way through to 2004 -- actually, I'm
       sorry, if you go up to the top of the page you will see
       there is more.  But it's from January 2001 all the way
       through to December 2004 and you will see at the bottom
       there's a figure of 27,005 calls.  Mr Parker's evidence
       is that's the number of incidents/calls that came into
       the third line support during the years that you were
       there.  Are you in a position to dispute that?
   A.  I would say that's probably the number of recorded calls
       that came in, yes.
   Q.  I'm grateful.  If we could go to the next tab, which is
       "RRP live PEAKs out of SSC", here is the calculation,
       again from January 2001 all the way through
       to December 2004, of the total number of calls that were
       passed out of third line support and we have already
       established that a proportion of those calls would have
       gone elsewhere?
   A.  Mm-hm.
   Q.  Some of them would have gone to fourth line support and
       I believe you have already accepted that a small
       proportion of those going to fourth line support would
       have gone to fourth line support because they required
       software fixes?
   A.  Yes.
   Q.  Very good.
           Now, the total figure that Mr Parker has analysed
       comes to 3,764 calls.  Do you see that?
   A.  Yes.
   Q.  Are you in a position to dispute that calculation?
   A.  I can't dispute that, no.
   Q.  No.  And I didn't expect you to.
           So what that means is that of the calls coming into
       third line support while you were there, 14% of them
       went elsewhere and of that 14% only a subset would have
       gone to fourth line support.  Do you accept that?
   A.  Yes.
   Q.  And of that subset only a small subset would have
       represented calls relating to software errors requiring
       fixing by fourth line support.  Do you accept that?
   A.  Does this figure, 3,764 -- is that -- and the other
       figure earlier on, is that calls that have come in, gone
       out then come back to us and gone out again somewhere
       else, or is it just ones that have come in, gone all
       round the houses, as it were, and then gone out for
       final resolution?
   Q.  My understanding is it is the latter but I had better
       check.  Yes, Mr Parker is nodding.  Are you in
       a position to challenge that?
   A.  No.
   Q.  Very good.
           So what that indicates then is summarised by
       Mr Parker in paragraph 32 on page {E2/11/9} and
       Mr Parker says:
           "From the SSC, only a tiny proportion of incidents
       were escalated to the 4th line support team.  It follows
       that only a tiny fraction of incidents raised actually
       needed to be looked at by the only team who might
       potentially effect changes in software."
           Do you accept that conclusion?
   A.  Which paragraph are we looking at now, sorry?
   Q.  Paragraph 32, Mr Roll.
   A.  32, sorry.
   Q.  The first sentence is right, isn't it?
   A.  When you take it as an average then yes I suppose so,
       but the system evolved from -- I started there in 2000
       I think it was and left in August 2004.  The nature of
       the work changed over time and the ones that stick in my
       mind are the ones where there was sort of the
       fire-fighting efforts where it was -- there were
       difficult periods where there were software issues,
       so ...
   Q.  Yes, Mr Roll.  We could by all means go back to the
       graphs because you will have seen from the live PEAKs
       into and out of the SSC it actually gave monthly
       figures, but what I'm suggesting to you, Mr Roll, is
       that -- although you may well have been busy on all
       sorts of things, the fact of the matter is that software
       problems requiring a software fix represented a tiny
       fraction of the work that was handled by the SSC third
       line support.
   A.  Yes.
   Q.  I'm grateful.  Now, you don't say anything about your
       seniority, although you have already expressed some
       modesty in that respect.  We have established that you
       weren't part of fourth line.  Is it fair to say that you
       were described as a SSC support consultant, was that
       your job title?
   A.  I think I was a product specialist.
   Q.  I'm sorry?
   A.  I believe my job title was product specialist.
   Q.  You were a product specialist and was that the junior
       level of people working --
   A.  It was the generic level.  I believe most people there
       were product specialists.
   Q.  And I think we have established that is it 30 people
       worked in --
   A.  Approximately.
   Q.  -- the SSC?  Of that 30 people how many people were at
       your level?
   A.  When I worked there I -- it was two or three people were
       senior levels, I think Mr Parker was one, a couple of
       others, then there was Mik Peach and I believe the rest
       of us -- 25 or so.
   Q.  The figures I would like to ask suggest to you, Mr Roll,
       are about 25 people had the junior level and there were
       about five people who were true specialists who I think
       you fairly described as specialist earlier on in your
       evidence.  Would you accept that that might be the case?
   A.  Junior level seems to -- it doesn't fully explain the
       complexity of the system or the knowledge of the system
       required, but yes, I suppose ...
   Q.  Well, let's agree on ordinary -- I'm not seeking to cast
       any imputations and it is right that you should -- words
       do have implications.  So the ordinary level was your
       level --
   A.  Base level.
   Q.  -- and there was this perhaps five or so people above --
       the senior people above you and what did they do?  Did
       they do the more challenging work?
   A.  Their experience and knowledge was more focused, shall
       we say.  Some people had a very good -- even I suppose
       some of the junior level had a very in-depth knowledge
       of the accounting side of things, so certain aspects
       would be passed over to them.  A couple were very, very
       clued up on UNIX and that side of things so the actual
       servers, any problems with those, if it was involved
       problems would go over to them.  So that was where their
       knowledge lay in specific areas, very in-depth very
       specific knowledge of those particular areas.  Does that
       explain the question?
   Q.  I'm really trying to grope towards what the five true
       specialists, what kind of work they were doing.  They
       were doing the more challenging kind of work, weren't
   A.  I find it difficult to answer that from what I remember
       of the way we worked.  There were some areas where some
       of the senior people where I would perhaps have been
       more -- had more experience because of previous work and
       previous programming that I would have been better off
       and had more knowledge than they would have done, but in
       other areas then they would be far superior to me.
   Q.  Well, let me suggest one area where they would have been
       far superior to you, Mr Roll, or at least it was
       perceived within the organisation -- I'm not trying to
       have a debate with you about your own perceptions of
       yourself, that would be completely unfair, but the
       perception at the SSC was that were there software
       errors or potential software errors that required large
       amounts of code to be examined, the people who would
       generally do that examination would be those five
       people, people like Mr Parker.  Would you accept that?
   A.  Generally I suppose, yes, although --
   Q.  And it would be relatively rare for someone outside that
       core specialist team to be doing that kind of work --
       not impossible but relatively rare, yes?
   A.  Several of us looked at the code on occasions.  I was
       a C programmer in that and other languages before
       I moved to Fujitsu so there were areas of knowledge we
       had from previous areas, but I suppose most of it would
       then be passed on to those people.
   Q.  I see.  So would you accept from me then that generally
       it would be them that would look at lots of code but
       occasionally there might be occasions when someone else
       would look at lots of code?
   A.  From a PINICL perspective yes, but some of us looked at
       code more often, just out of interest.
   Q.  Well, I will ask you about that in a while.
           Now, the last thing I'm going to be doing in this
       cross-examination is suggesting that you didn't have
       a responsible position and a responsible job.  You were
       well trained, weren't you?
   A.  I was well trained by Fujitsu, yes.
   Q.  Is it the case that there was a mandatory six month
       training process you had to go through at Fujitsu before
       you were allowed unsupervised access to the live system?
   A.  I don't fully recall that.
   Q.  And would it also be fair to say that the work that you
       did involved tasks that were much easier and were much
       more mundane -- I don't mean to be insulting -- than the
       kind of tasks that the senior people within the third
       line of support did?
   A.  I think one of the comments Mr Parker made in his
       statement at some point was that some things stick in
       the mind more than others and over time one tends to
       forget the mundane aspects of the job and tends to
       remember more the aspects that were not mundane, so my
       recollection is that -- as I put in my statement --
       I spent sort of 70% of the time doing not mundane tasks,
       as it were, and only 30%, but that could be wrong, it
       could be my perception of events from 15/17 years ago.
   Q.  That's very fair, Mr Roll, and I'm grateful for that.
           Can we look at some PEAKs which indicate the kind of
       work that you were doing and I'm not suggesting, by the
       way, that these PEAKs are representative of all the work
       that you did.  But if we can go to {F/94.1} please.
       That's a PEAK that's dated 26 March 2001 so it is quite
       early on in your tenure at the SSC.  Could I ask you
       just to -- well, let's -- it's halfway down the first
       box we've got "Information: long calls detected on
       Bootle ISDN66" and then there is a series of times
       relating to particular branches, particular FADs.  Are
       you able to explain what that's about?  What that
       problem was?
   A.  I think from what I remember calls were only supposed to
       last a certain time.  If they lasted more than so many
       minutes then it would flag up an error.  I think that's
       what that says but that's --
   Q.  I see, and then you will see several boxes down under
       the heading "Graham Critchley", do you see that?
   A.  Yes.
   Q.  And then the next box down:
           "F response: lines checked, no longer up.  All long
       calls cleared.  KEL updated."
           And then there is a reference to a KEL.
           Can you describe what you were doing in relation to
   A.  No, I can't remember.
   Q.  You can't remember.
           Then you would have been -- it would have been your
       final response.  The fact that your name is at the
       bottom indicates that you are the final responder for
       this PEAK, yes?
   A.  Yes.
   Q.  And generally speaking would it be right to say that the
       final responder is the person who has ownership of the
       conclusion of the process?
   A.  Yes.
   Q.  And the resolution and usually knows more about the
       problem than anybody else, would that be fair?
   A.  Not necessarily knows more about the problem, no, but
       would have ownership of the problem.
   Q.  And its resolution?
   A.  Yes.
   Q.  And if we could move on to {F/164.01} please.  This is
       a PEAK -- it is now a couple of years, perhaps three
       years into your tenure.  It is dated 28 November 2003
       and at the top on the first box we've got at 15/10/03,
       11.17 and then there is a long acronym.  Could you tell
       the court what OBC stands for?
   A.  I don't remember.
   Q.  "Engineer to site 21/11/03 am to remove 2 counters and
       pin pads from site and return to Triage Services.
       Engineer to ensure EOD marker is received prior to
           So an engineer is going into a branch to remove some
   A.  Looks like the branch has closed.
   Q.  It looks like the branch is closing.
           Then if we go to the next box we've got
       Barbara Longley -- who was Barbara Longley, do you
       remember her?
   A.  I think she was the lady who was the administrative
       person responsible for allocating calls.
   Q.  And then several lines down, "the call summary is now"
       and there is a FAD number "check for marooned
       transactions".  Could you explain what the marooned
       transactions are?
   A.  I can't really remember but I think you had an end of
       date marker and if anything had been -- any transactions
       that were done after the last end of day marker -- so if
       the branch had closed say at lunchtime, anything done in
       the morning there wouldn't have been an end of data
       marker left so I believe they may not have been
       harvested properly and I think that was ...
   Q.  Well, can I suggest to you that what's happening here is
       that there's a basic kit removal operation and marooned
       transactions are transactions that might be located on
       the machines but are trapped in -- they are stuck in the
       machines, possibly because of a hardware error.  Would
       that be right?
   A.  Possibly.
   Q.  I see.  And you will see halfway down the same box:
           "Pre-scan: assigning call to Richard Roll in EDSC."
           Do you remember what EDSC stands for?
   A.  No, I don't.
   Q.  But it is assigned to you and then if we go to the next
       box down your response is:
           "There are no marooned transactions; EODs and cash
       accounts harvested okay."
           So what work would you have done there to result in
       that report?
   A.  I don't know.
   Q.  You can't recall?
   A.  I don't remember.
   Q.  I see.  Well, let's try the next one {F/193.1} please.
       This is dated 26 February 2004.  This is the year you
       left and the first big box with big text in, another OBC
       exercise of some sort, "planned short notice close", so
       it looks as if a branch is closing down:
           "ROMEC Engineer to site 18/02/2004 AM to remove 1
       counter and pin pad from site and return to triage
       Services.  Engineer to ensure EOD marker is received
       prior to decommissioning."
           So an engineer is going to a branch to close it down
       and remove the equipment, so similar to what we saw
           And then the next box down, about a third of the way
       down there's a line which reads:
           "Attended site [18 February].  SMC John confirmed.
       Synchronisation ... ROMEC work complete."
           Who were ROMEC, do you remember?
   A.  No.
   Q.  And two lines down:
           "Information: please check for last CA and potential
       marooned transactions."
           Do you know what "last CA" means?
   A.  Last cash account I think.
   Q.  And potential marooned transactions, again is someone
       seeking to check whether there were any transactions
       stuck on a machine that haven't been harvested and sent
       elsewhere, is that right?
   A.  Yes.
   Q.  And then if we go down the page to the bottom of the
       page, right at the bottom it is assigned to you, so you
       are still doing this kind of work amongst many other
       kinds of work --
   A.  Yes.
   Q.  -- in 2004 and over the page {F/193.1/2} about halfway
       down, 19 February at 11.8, user Richard Roll:
           "Final cash account harvested correctly, there are
       no marooned transactions.  No further action required."
           Now, is it fair to say that a significant portion of
       your work did involve supporting engineers doing that
       kind of work?  Would that be right?
   A.  I can't remember what percentage was what.  It would
       have varied over time.  Certainly for the last year
       after I had handed my notice in, I'm not -- I can't
       remember if there was a change in my workload or not.
   Q.  Well, let's go to your second witness statement that you
       made in January of this year.  It is at {E1/10}.  You
       have already alluded to it.  I would like to go to
       page 8 {E1/10/8} which is paragraph 25.  Perhaps you
       could just read it quietly to yourself.
   A.  Yes.
   Q.  About a third of the way down you say:
           "However, if it is helpful to the court for me to at
       least clarify my role, then I would add the following
       short points: 1, whilst my workload did involve some
       support to engineers opening and closing branches,
       I would estimate that this made up only 30% of my work,
       and the majority of my workload (estimate 70%) involved
       looking for faults on data stores, preparing reports for
       the manager as a result of reports of problems with
       Horizon experienced by the estate ..."
           And then you say:
           "... as part of this role, at least some of the time
       I was involved in examining source code ..."
   A.  Yes.
   Q.  Now, first of all, when you say "looking for faults on
       data stores", what do you mean?  Would that include
       looking for marooned transactions on counters, on the
       message stores?
   A.  It could do, yes.
   Q.  I see.  And would it include checking to see whether
       message -- whether counters, whether the machines are
       locked when they should be unlocked, that kind of --
       looking for problems in the configuration data for
       example in machines which prevent them from working
       properly?  Would you include that as looking for faults
       on data stores?  Because it does seem to be what those
       words mean, Mr Roll?
   A.  I find -- I can't remember any specific details of that
       sort of depth of work that I did.  What I believe was
       that I said most of my work was of a more technical
       nature, not supporting the engineers, but that's my
       recollection and that's the best I can give you.
   Q.  Faults in data or data stores, that doesn't mean looking
       for bugs in software, does it?
   A.  It's not technically looking for a bug in the code, more
       a bug in the data, a corruption in the data.
   Q.  Yes.  So that's a very useful distinction actually.
       A difference between errors in code, bugs in -- what
       layman would probably call bugs and then data
   A.  Yes.
   Q.  And data corruption is a vast sort of panoply of
       problems.  It could include configuration problems, you
       could have data which has some fields that are missing,
       you could have machines that are locked because they've
       got the wrong binary setting -- I'm not putting this
       very technically, I hope you will forgive me.  It would
       will include a vast array of problems, many of which
       would be hardware related, would that be right?
   A.  No, not necessarily.  The data ... it could -- it's not
       necessarily the hardware that's corrupted it.  The data
       becomes corrupt.  Data is not hardware or software, it's
       the data.
   Q.  Yes, when I say "many of which could be hardware
       related" I'm actually trying to put that very point to
       you, Mr Roll, which is that if you have a data
       corruption, it may be the result of a software bug --
   A.  Yes.
   Q.  -- but it may well be the result of something completely
       different, there are a wide array of things, including
       hardware issues, would you accept that?
   A.  Yes.
   Q.  So when you describe 70 per cent of your work as
       including looking for faults on data stores, you are not
       saying, are you, that that work was looking for software
       bugs?  That's not what you mean?
   A.  Not 70% of it, no.
   Q.  I'm grateful.  And then you say:
           "... preparing reports for the manager as a result
       of reports of problems with Horizon experienced by the
       estate ..."
           First of all, who is the manager in that sentence?
   A.  Mik Peach.
   Q.  I see.  So what you were doing, he would appoint you to
       provide him with reports, would he?
   A.  He did, yes.
   Q.  And those reports -- he would do that because he himself
       had received reports of one kind or another, is that
   A.  He -- Mik Peach was engaged to another lady who worked
       in Fujitsu who was of a similar or maybe senior level to
       him and she occasionally asked him for favours and he
       would ask me to write reports and generate reports to
       give to her.
   Q.  You describe them as being "as a result of reports of
       problems with Horizon experienced by the estate".  What
       are these reports of problems with Horizon experienced
       by the estate?
   A.  I don't remember the details.
   Q.  Are we talking about standard routine reports that are
       generated automatically?
   A.  No.
   Q.  So we're talking about specific reports, what, by human
   A.  Specific reports where I would look at the various
       reports that had already been generated, collate the
       data and extract specific data for him.
   Q.  I see.  So there were people producing reports and then
       Mr Peach would then ask you to do reports collating the
       information that was in those underlying reports; is
       that what you are describing?
   A.  Some of the reports were systems generated, some were
       generated by other people.
   Q.  And systems generated -- I would like to investigate
       that a little bit.  This also applies to the looking for
       faults on data stores that you describe.  A lot of the
       work you did in looking for faults on data stores was
       related to system reports that -- there are a series,
       a vast series actually, of reports that are
       automatically generated by the system as to its own
       performance just to make sure it is working, lots and
       lots of automated reports, and when you were looking for
       faults on data stores you would be using those reports
       to look for problems which you would then be looking to
       see whether you could see if there was a problem and if
       so what should be done about it?
   A.  Okay.  Some of the reports were looking for
       communications problems, so long calls.  Others would be
       looking for errors in the data stores.  The data store
       was harvested every day and the information would then
       be processed overnight, so there would be all the
       banking transactions from all the counters would be
       extracted and batched up and then sent off to different
       banks and every process along there would generate
       a report as well.  So it wasn't just the data stores, it
       was the whole system was generating reports.
   Q.  There were an enormous --
   A.  We would have access to all those reports, sorry.
   Q.  I'm sorry, I didn't hear you that last -- that's my
       fault, I talked over you and now I'm saying I can't hear
       you.  I'm sorry, Mr Roll.
           Yes, I see.
           So you would agree with me, would you, that there
       were a large number of automatic reports that were
       continually produced by the Horizon system --
   A.  Yes.
   Q.  -- to monitor its performance and to look for problems?
   A.  Yes.
   Q.  And that when those reports generated some hits when
       potential problems were identified, then people at your
       level would go and look at them to see whether there was
       anything that needed to be done about them?
   A.  Yes.  As time progressed we wrote more and more reports
       ourselves -- or programmes ourselves to monitor the
       system, to alert us to any potential problems.
   Q.  And then perhaps I could ask you to go back to
       Mr Parker's statement, {E2/11/9}.  If I could pick it up
       at paragraph 34.  Mr Parker says:
           "As noted above, the SSC team provided both 2nd and
       3rd line support.  As with any mix of people, there are
       (and were) various levels of talent within SSC.  Mr Roll
       was primarily used in operational business change (OBC),
       which involved supporting the engineers who were opening
       and closing branches and increasing and decreasing the
       number of counters in branches."
           Stopping there, does that remind you as to what OBC
       stands for?
   A.  It does, now.
   Q.  And would you agree that you were primarily involved in
       that activity?
   A.  That's not how I remember it, no.
   Q.  Could I suggest to you that it is quite possible that
       because of the process you yourself described a few
       minutes ago, that one tends to remember only the more
       interesting things --
   A.  It is, it is possible.
   Q.  -- it is actually possible that that did represent your
       primary set of jobs?
   A.  Yes, it could be.
   Q.  I'm grateful.  He says:
           "Mr Roll would also have been regularly correcting
       the application environment after engineers had replaced
       failed counter hardware and clearing temporary files to
       increase disc space."
           Is that true?
   A.  It's possible.  I don't remember it, but it's possible.
   Q.  "This could fairly be described as 2nd line work and it
       was done by the SSC because it required a higher level
       of access to the system than other support teams had."
           Would you accept that?
   A.  It could be considered that.
   Q.  And then paragraph 35:
           "Some members of the 3rd line support group
       identified the need for software fixes via source code
       examination and would pass this on to the 4th line team
       for a code fix to be written.  Mr Roll did not play any
       significant part in this and was not involved in any
       extensive source code examination.  An application code
       fix would not be written by anyone in the 3rd line team
       and he was not involved in the provision of 4th line
           Could I suggest to you, Mr Roll, that you did not
       play any significant part -- I'm not saying you played
       no part, but you did not play a significant part in that
       process and you weren't generally involved in extensive
       source code examination?
   A.  That is correct, not generally involved, but I was
       involved in some source code examination.
   Q.  On occasion?
   A.  On occasion.
   Q.  And then would you agree that an application code fix,
       that's a software fix to correct what I would call
       a bug, would not be written by anyone in the 3rd line
       team; that's right, isn't it?
   A.  Yes, we would identify a problem but the software
       development team would write it.
   Q.  It's the fourth line people that fix them and I think
       you would agree that from Mr Parker says here is that
       you were not involved in the provision of fourth line
   MR JUSTICE FRASER:  Is that right?
   A.  Fourth line support, from that way -- I suppose yes, you
       could say that.  The way I was looking at fourth line
       support was writing code that would monitor the systems
       and generate errors -- generate alerts if there was
       a potential error, so from the way that's put then yes,
       you could say that.
   MR DE GARR ROBINSON:  You weren't involved in the support
       that was provided by the fourth line?
   A.  No.
   Q.  Thank you.
           So would it be fair to say that the reading of
       copious amounts of code was relatively rare for the
       third line of support and when it did happen it was
       generally done by the senior people that we have
   A.  Yes, I would think so.
   Q.  And although you did it on occasion, it was even rarer
       for you; would that be --
   A.  Yes.
   Q.  I'm grateful.
           Then if we move on to paragraph 37 {E2/11/10}.
       Perhaps I could ask you to read paragraph 37.
   A.  Sorry, read out loud or ..?
   Q.  No, no, just read it to yourself.  I'm so sorry.
           Subject to one point which I'm about to raise with
       you, Mr Parker says here:
           "Moreover, Mr Roll was not working at a level where
       he would be required to review code."
           Mr Roll, I'm not suggesting that you never reviewed
       code, sometimes you did, but subject to those occasional
       times would you accept what Mr Parker says in
       paragraph 37?
   A.  It's again what we said earlier.  My recollection is
       different to this but it's possible that predominantly
       I did do the workaround.  It's a long time ago.
   MR JUSTICE FRASER:  What about the sentences that say about
       the role of the SSC, do you agree with them?  The SSC
       had access to view, but couldn't amend, source code?
   A.  Yes, I agree with that.
   MR JUSTICE FRASER:  That's correct.  And the access which is
       then described in the second sentence was rarely used,
       or do you not know?
   A.  I looked at the source code on several occasions, but
       over a four year period that is rarely used, so ...
   MR DE GARR ROBINSON:  Well then, Mr Roll, if I could ask you
       to go back to your first witness statement, it is E1,
       tab 7.  If we could pick it up at paragraph 6 {E1/7/1}
       you say:
           "My role involved receiving issues or problems being
       experienced by people working within Post Office
       branches.  In many cases, this involved me dealing
       directly with subpostmasters and others employed in
       Post Office branches."
           And then in paragraph 7 you say:
           "By way of example of the type of issue that I would
       deal with, if a financial discrepancy had arisen in
       a branch (eg a 'shortfall' of £5,000) then I would need
       to work sequentially through all transactions over the
       relevant period and also work through thousands of lines
       of computer coding.  Software programmes were written by
       us to strip out irrelevant data to enable us to more
       easily locate the error."
           Given the discussion we have just been having,
       Mr Roll, would you accept that perhaps with the benefit
       of hindsight paragraph 7 might give a slightly
       misleading impression, I'm sure unintentionally.  It
       gives, I would suggest, the impression that this is
       something you did regularly, that when dealing with
       problems arising in branches you would regularly work
       through thousands of lines of computer coding, but
       that's not actually the case, is it?
   A.  That is not the case, no, and that is -- I must agree
       that is what that suggests.
   Q.  That's fair.  That wasn't your intention to suggest that
       and it may be I'm reading too much into it.  Is that
       what you are saying?
   A.  My recollection was that I did work through computer
       coding of the programmes, but now I'm not so sure.
   Q.  I'm very grateful, Mr Roll.
           Then if we could go back to {E2/11}, Mr Parker's
       statement.  If we could pick it up at page 10
       {E2/11/10}, in paragraph 39 Mr Parker describes how:
           "When an incident is resolved, the SSC team member
       (or technician as they are sometimes called) types
       a summary of the incident (known as a final response)
       and allocates a response code to the incident in order
       to classify it."
           Pausing there, do you remember that that was how the
       SSC worked?
   A.  No, I don't remember that.
   Q.  You do remember the fact that someone would be
       responsible for producing the final response on any
       PEAK, yes?
   A.  Yes.
   Q.  And do you remember that that person would allocate
       a code to the PEAK reflecting his or her judgment as to
       the nature of the problem that had been dealt with?
   A.  I know that that's what happened because I have seen the
       documents here, but I don't remember that.
   Q.  Oh, well, it might be that the questions I was about to
       ask you could go quicker but I will try anyway.
           "While guidance is provided on when to use each
       response code ... allocation is the subjective view of
       the technician closing the incident and there is no
       re-examination of the response codes later to ensure
           Would I be right in thinking that you are not really
       in a position to comment at this remove of time?
   A.  No.
   Q.  Then Mr Parker says in paragraph 40:
           "With that in mind, the final response codes that
       were allocated to incidents ... reported to SSC between
       1 January 2010 and 31 December 2004 were as follows:-
           "Known issue/workaround - 35.3%;
           "Admin - 27.5%;
           "Reconciliation - 15.7%;
           "Potential user error - 10.9%;
           "Potential software error - 8.3%; and
           "Hardware error - 1.2%."
           Are those terms, those characterisations, known
       issue/workarounds, admin, reconciliation, are they
       familiar to you?  Would you be able to -- do you have
       a sense of what these things mean?
   A.  Not from my time there, I can't remember them.
   Q.  You can't remember.  Well, let me ask you this then.
       You will see in paragraph 40.5 he suggests that if you
       review the PEAKs for the relevant period the PEAKs that
       qualified as potential software errors represented 8.3%
       of the work that was done by the SSC.  Would you be in
       a position to challenge that or accept it?
   A.  No, neither.  I wouldn't --
   Q.  Neither?  You just don't know.  Fair enough.
           Then Mr Parker says in paragraph 41:
           "A major part of 1st line's raison d'etre is to deal
       with user error and therefore the percentage of issues
       attributable to user error would be much higher at
       1st line."
           Would you agree with that?
   A.  Yes.
   Q.  And 41.2 {E2/11/11}:
           "Very few hardware incidents reached the SSC because
       they were the preserve of the HSD (ie they were
       relatively easy to spot and therefore filtered out by
       1st line support)."
           Would you agree with that?
   A.  Yes.
   Q.  41.3:
           "8.3% of calls to the SSC ... are attributed to
       potential software errors over these four years.  This
       includes duplicates ..."
           Would that be right, do you recall?
   A.  No.
   Q.  Would you be in a position to dispute that --
   A.  No.
   Q.  -- the 8.3% -- the PEAK analysis would include PEAKs
       that were essentially duplicates of other PEAKs?
   A.  No, I couldn't -- I couldn't --
   Q.  You couldn't remember?
   A.  No.
   Q.  So he also says this:
           "... does not provide any clarity on the
       significance of the error corrected."
           I suppose that's a statement of the obvious,
       isn't it, you would accept that?  The 8.3% could include
       trivial problems, could involve major problems?
   A.  Yes.
   Q.  "Many software errors, particularly in a new product,
       were insignificant, such as correcting capitalisation in
       printed output."
           Would you agree with that?
   A.  Yes, I assume so.  I don't know if anything that trivial
       would have been passed through to us, or maybe it would,
       I can't remember.
   Q.  Well, if we could look at your own -- my Lord, I see
       that it is 25 past 3.  There was a break earlier, but
       would your Lordship --
   MR JUSTICE FRASER:  Well, I assume -- are you going to go
       what until about quarter past 4, half past 4?
   MR DE GARR ROBINSON:  If I could go to half past 4 I would
       be grateful.
   MR JUSTICE FRASER:  Let's have a break now for the shorthand
           Mr Roll, you are in the middle of giving your
       evidence.  Please don't talk to anyone in the short
       break or indeed overnight about your evidence, because
       you are being cross-examined.  We will have a break
       until 25 to for the shorthand writers, and if everybody
       could come back for that time that would be good.  Thank
       you very much.
   (3.28 pm)
                          (Short Break)
   (3.38 pm)
   MR DE GARR ROBINSON:  My Lord, thank you.
           Could we go to {F/1839} please and once you've got
       it up I would like the "RRP live PEAK by category" tab
       again, if that's possible.
   MR JUSTICE FRASER:  Do you want in or out?
   MR DE GARR ROBINSON:  Could you keep going, I'm so sorry.
       Yes, that's the one.
           So, Mr Roll, can you see that there's a table
       here -- if you could perhaps bring the table up so we
       could see the whole of it because we are missing the
       first nine rows, I think.
           This is a table which Mr Parker describes in his
       witness statement.  It is an analysis of the PEAKs that
       were produced during the four years 2001 to 2004 and it
       is an analysis of the codes that were applied by the
       final responder to the relevant PEAKs and do you see,
       Mr Roll, that there are a number of -- there's
       a column on A which is "Closure category" and these are
       the codes, the codes that were used at the time that you
       very fairly said you don't remember, it being so long
       ago, and all the numbers going down from zero which is
       "Deleted", all the way up to 200.  Then there is a brief
       description of these codes and then there is the
       column C is the overall count of total number of PEAKs
       with these particular codes and you will see it comes to
       27,002 and I should -- it is worth mentioning, my Lord,
       Mr Parker says there are 27,005.  This says 27,002,
       there are three missing, I'm afraid I'm not in
       a position to tell anyone why.
   MR JUSTICE FRASER:  They might be in cell C2 which has been
       deleted.  It might not count.
           Can we just click on cell 34C -- I mean it is not
       a big point so it is not worth ... no, sometimes it
       shows the formula.
           All right, it doesn't matter.
   MR DE GARR ROBINSON:  I'm afraid it was posited only
       a couple of hours ago and I'm not in a position to --
   MR JUSTICE FRASER:  I don't think the difference between
       27,002 and 27,005 is going to make any difference.
   MR DE GARR ROBINSON:  I don't either.
           So you will see, Mr Roll, just to explain it,
       there's the overall analysis of the codes, the
       distribution of the PEAKs between these different codes
       and then there is a -- if you go -- you can skip D for
       the moment, but E, F, G, H, I, J and K, these are sort
       of broader categories which are intended to encompass
       different groups of code, so Mr Parker has chosen
       "Potential software error" then "Potential user error",
       "Hardware", "Admin", "Known issue/work round",
       "Reconciliation" and "Documentation", do you see those
       broad categories?
   A.  Yes.
   Q.  Just to explain how it works -- it is not nearly as
       complicated as it looks.  If you look at "Potential
       software error" you go down and you see there is a 1 on
       row 12.  That corresponds to closure category 59:
           "Ref data fix available to call logger."
           That's the official code, or rather a brief summary
       of code 59 and it shows that there are some hits in that
       code category during the four years that you are there.
       Do you understand how it works?
   A.  Yes.
   Q.  Then next one down, another potential software error is
           "Software fix available to call logger."
           And the third one down is 61 "Build fix available to
       call logger".  Do these phrases, "Reference data fix
       available to call logger", "Software fix available to
       call logger", "Build fix available to call logger", do
       those terms ring any bells?
   A.  No.
   Q.  Or is it really just to -- would I be wasting valuable
       time trying to --
   A.  I'm afraid I can't remember any of those.
   Q.  That's fair and I won't -- and then just to complete the
       picture, row 24 there's a code 74 which is also treated
       as a potential software error and this is "Fixed at
       future release" but given what you have just said
       perhaps I will leave it at the fact that you are not in
       a position to challenge these categorisations, are you?
   A.  No.
   Q.  And then we've got "Known issue/workround" in column I.
       The first one, the first hit is at 64, "publish known
       error", that sounds quite logical, so if there's a known
       error log already in existence that's just taking it and
       saying "This is the KEL that applies to this problem".
       And the next one down is "Unpublished known error" and
       the next one down is "Solicited known error" and the
       next one down is "Avoidance action supplied".  Again,
       would I be right in thinking it is probably not
       a valuable use of time to ask you about those
   A.  No.
   Q.  Okay.  Well, this is Mr Parker's categorisation and you
       will see that he distinguishes between potential
       software error and known issue/work-arounds and if we go
       down to the bottom of the table, looking at the overall
       figures of the 27,002, we have 2,252 which are treated
       as potential software errors, that's 8.3% and I think we
       have already discussed that figure and I believe you
       said you are not in a position to dispute that
       calculation, is that right?
   A.  That's correct.
   Q.  And then if we go over to "Known issue/work round" there
       are 9,545 of those which represents 35%, 35.3% of the
       work being undertaken by the SSC while you were there.
       Again, would I be right in thinking you are not in
       a position to dispute that?
   A.  Yes.
   Q.  Thank you.
           Now, staying with this document I would like to go
       to a different tab please.  Could we go to the
       "Final responses" tab.  Now, this, Mr Roll, is an
       analysis of the PEAKs for which you were the final
   A.  Right.
   Q.  And you will recall that this is addressed in
       Mr Parker's witness statement.  You will see that there
       are 915 PEAKs in which you are the final responder while
       you were there in your 44 months when you were at
       Fujitsu and there's a breakdown of which particular
       categories you had applied to the relevant PEAKs when
       you closed them.
   A.  Right.
   Q.  Do you see that?  And if we could go back please to "RRP
       live PEAK by category" which I think is to the right,
       you will see that as well as an overall analysis of the
       work done by the SSC there's also an analysis of the
       would, that you did, "RR", this is in the pale -- is it
       beige?  When I was a child I think I would have called
       it fawn but we have to say beige now.  There is a beige
       line and it has your 915 PEAKs and you will see
       "Potential software error", the total for the SSC is
       8.3% but do you see that your total comes out at 3.2%?
   A.  Mm-hm.
   Q.  Would you be in a position to challenge that percentage?
   A.  No.
   Q.  Is it fair to say that it would reflect the nature of
       the work that you did because you weren't one of the
       senior specialists?
   A.  It's possible, yes.
   Q.  And if we go along to "Known issue/work round", which is
       column I, as I say we have 35.3% for the SSC but we've
       got 61%, nearly 62% for you.  Again, would you be in
       a position to challenge that allocation?
   A.  Not on the figures here.  What I can say is that in some
       instances more than one person worked on a snag at
       a time, but their involvement may not be included in
       this and also that over time my workload would have
       shifted and ... well, I --
   Q.  I see what you are saying.  The point about this doesn't
       cover all the work you did because you would have worked
       on other PEAKs for which you were not the final
       responder, could I just suggest to you that that may be
       true, I'm sure that is true, but it's also true,
       isn't it, that the PEAKs on which you are the final
       responder are representative of how you spent your time,
       the proportion of the work that you did as allocated
       between different -- it's unlikely that you would be
       a final responder in categories which didn't represent,
       you know, a fair allocation of the work you were
       actually doing at any particular time, would it?
   A.  It's interesting because this isn't how I remember it,
       but that's what the figures say so ...
   Q.  But just if I could put my point again, it's a small
       point and I will move on.  You say, fairly, that these
       are just -- assuming these figures are right and I'm not
       demanding that you accept them --
   A.  I -- yes.
   Q.  If these figures are right, these just represent the
       PEAKs in relation to which you were the final responder,
       you were the one who had ownership -- if I can use that
       horrible cliche -- of the resolution of the problem and
       the closure of the PEAK.  It's true that you would have
       worked on other PEAKs along the way, you would have
       been, you know, allocated the job and then it would pass
       to someone else and so on, it would have passed through
       your hands, but wouldn't it be right to say that the
       PEAKs where the music stopped -- the PEAKs that landed
       in your lap as the final responder would be a fair --
       a representative reflection of the nature of the work
       that you did as between different categories of work?
   A.  Yes, I should think so.  There were probably some that
       I got to start with that I then passed on to, as you
       said earlier on, more qualified personnel within the
   Q.  I see.
   MR JUSTICE FRASER:  Mr De Garr Robinson, just so that I can
       follow one of the tabs on this document can we go back
       to the -- I just have a question for you about the
       entry.  Can we go back to the "Final responses" tab
       please.  Do you see there from row 21 downwards on the
       left-hand side the category is given but for some reason
       there isn't a description on the left-hand side.  Is
       that just because there isn't enough room in that cell
       and I should look across to column D for the
   MR DE GARR ROBINSON:  Are we looking at column --
   MR JUSTICE FRASER:  I mean look, for example, at row 21
       which is "As category 70", row 13 category 70 then says
       "Avoidance action supplied" but for some reason it
       doesn't on row 21.  That might not be an issue because
       it has avoidance action in column D.
   MR DE GARR ROBINSON:  Let me take instructions.  I hadn't
       noticed that and I will take instructions.
   MR JUSTICE FRASER:  It is not important for now, just tell
       me in the morning.
   MR DE GARR ROBINSON:  I can tell your Lordship one thing
       your Lordship may find helpful is that there is
       a document which gives an indication as to what these
       codes actually mean.  It is from a later period in 2011
       but I'm instructed that it is broadly accurate and for
       your Lordship's note, if it your Lordship is interested
       as to what these codes actually mean, it is {F/823} and
       the analysis of the codes starts at page 23 of that
       {F/823/23}.  I'm not suggesting -- there are some codes
       that are omitted but if your Lordship would like to know
       more about those codes, it is actually -- I was going to
       take Mr Roll to them but given his --
   MR JUSTICE FRASER:  I imagined you probably would.  I was
       just curious about that column but it doesn't matter,
       please don't interrupt your flow.
   MR DE GARR ROBINSON:  I will take instructions overnight.
           I'm so sorry to keep you waiting.
   MR JUSTICE FRASER:  It is my fault for interrupting.
   MR DE GARR ROBINSON:  So, Mr Roll, with that analysis can we
       go back to your first statement please.  It is {E1/7/2}.
       It is paragraph 11 that I'm interested in.  You say:
           "I would estimate that there were over 30
       individuals working on the same floor as me ... in the
       Fujitsu building at Bracknell.  My recollection is that
       many of those individuals were involved in similar work
       to myself, or were involved in other Horizon related IT
       work.  I would describe much of the work being carried
       out as 'fire-fighting' coding problems in the Horizon
           For "coding problems" I take it you're talking about
       software bugs.  On the basis of the analysis of the
       evidence that we have just been through at slightly
       painful length, would you accept that it is possible
       that you may have rather exaggerated the position, no
       doubt unintentionally, in the last sentence of
       paragraph 11?
   A.  Certainly the way I remember it doesn't agree with the
       figures, so yes, it is possible that --
   Q.  Much of your work was not fire-fighting coding problems
       at all, was it, coding problems represented a tiny
       amount of your work?
   A.  On the evidence you have just supplied, that's correct.
   Q.  Nor did it occupy much of your colleagues' work either,
       if we look at the overall figures, 8.3% for software
       errors, would you accept that?  Even that -- it would be
       unfair to describe the process that was undertaken
       during that four-year period as fire-fighting coding
       errors by the SSC.  It's just not accurate, is it,
       Mr Roll?
   A.  From volume perhaps not.  Time-wise avoidance action
       would obviously take much less time than trying to find
       a software problem.  But the figures certainly
       contradict my recollections of it.
   Q.  That's fair.
           If we could move back to paragraph 9 {E1/7/2} you
           "We regularly identified issues with the computer
       coding in the Horizon system.  We would then flag those
       issues to the Fujitsu IT software developers."
           Mr Roll, you didn't regularly identify issues with
       computer coding that would be flagged for the fourth
       line support, did you?  That's just not consistent with
       the figures we have been looking at?
   A.  The way I remembered it was that we were as a department
       finding issues every week but I'm not sure if that is
       the case now.  That was my recollection.
   Q.  I'm grateful.  Then paragraph 8 -- this may have been
       carefully drafted, or it may actually have been
       carelessly drafted and I'm not attributing any motives
       to you, I'm genuinely not, but if you look at the last
       sentence of paragraph 8, you say:
           "If an error was referred to us then it was
       extremely unlikely to be due to a mistake made by
       a postmaster; the vast majority of errors I dealt with
       were due to coding errors or data corruption."
           Just to be clear, Mr Roll, you are not saying that
       the vast majority of errors you dealt with were coding
       errors, are you?
   A.  No.
   Q.  And when you say "or data corruption", that includes all
       sorts of problems --
   A.  All sorts of problems --
   Q.  -- that had nothing to do with software bugs?
   A.  Yes.
   Q.  I'm grateful.  Now, I would like to ask you about
       paragraph 10 of your witness statement {E1/7/2}.
           I'm sorry, Mr Roll, I'm told that the transcriber
       hasn't got the last answer.  Could I read my question
       again and then you can answer it again.  I think it may
       be my fault for having talked over you.  My question
       was: and when you say "or data corruption" that includes
       all sorts of problems that have nothing to do with
       software bugs?
   A.  Yes.
   Q.  I'm very grateful.
           Then paragraph 10 on page 2 {E1/7/2} of your witness
           "My recollection is that the software issues we were
       routinely encountering ..."
           Do you think there's a danger that by describing
       yourself as "routinely encountering" software errors,
       you may have unintentionally exaggerated the position?
   A.  I'm referencing here the whole department, not myself
       and it was my recollection that as a department we were
       routinely encountering software errors, which, as we
       have previously said, the figures don't support that but
       that was my recollection.
   Q.  That's very fair, Mr Roll, thank you.
           You say:
           "My recollection is that the software issues we were
       routinely encountering could and did cause financial
       discrepancies at branch level, including 'shortfalls'
       being incorrectly shown on the Horizon system."
           I think you will accept, won't you, that software
       issues did not routinely cause branch discrepancies?
       You are not saying that, are you?
   A.  By "software" there I mean software and data issues,
       corruption, et cetera.
   Q.  You're including data corruption issues and --
   A.  The whole thing, yes.
   Q.  I see.  So when you say "software issues" we should --
       would it be fair to say that whenever you refer to
       "software issues" we should probably read in the
       possibility of other data issues that aren't concerned
       with software bugs?
   A.  Yes.
   Q.  I'm grateful.
   MR JUSTICE FRASER:  By "software bugs" I think
       Mr De Garr Robinson is talking about code issues.
   A.  Yes.
   MR JUSTICE FRASER:  That's how you are putting it.  You are
       using in your statement, as far as I can tell,
       "software" in a rather more generic way, is that
   A.  I'm afraid so, yes.  I should have -- it has been a long
   MR JUSTICE FRASER:  That's my understanding from both other
       cases and today.  Coding is what in the computer
       industry they would describe as software issues, I think
       and that's the way you are putting your questions.
   MR DE GARR ROBINSON:  Yes and I'm suggesting that the fact
       that you have a data problem --
   MR JUSTICE FRASER:  I understand that.
   MR DE GARR ROBINSON:  -- bears no necessary relation to --
   MR JUSTICE FRASER:  I entirely understand all of those
       points.  Just so that you, Mr Roll, if you could specify
       in any of your answers when you are referring to coding
       using the word "code" then I will understand and
       Mr De Garr Robinson will understand what your answer
       relates to.
   A.  Yes, my Lord.
   MR JUSTICE FRASER:  All right.
   MR DE GARR ROBINSON:  So back to paragraph 10, I think you
       have -- I hope I'm not putting words into your mouth.
       Well, let me -- the software issues properly
       so-called -- that's unfair, that's suggesting I'm
       pulling rank on you or something, that's not my
           Software issues in my definition, bugs.
   A.  Bugs.
   Q.  Coding issues.  You are not saying that even the whole
       group of you, not just you, routinely encountered issues
       of that sort?  You're not saying that, are you?
   A.  We would -- a call would come into us with an error in
       the accounts or something like that basically.  We would
       then look for a cause.  And so unfortunately my
       terminology -- it has been a long time since I was in
       the industry.  So it could be the code or it could be
       the data corruption and we would have to try and find
       out which one it was.
   Q.  So you are not saying that you were routinely
       encountering coding issues, are you?
   A.  Bugs, no.
   Q.  I'm grateful.  And then you said these issues could and
       did cause financial discrepancies at branch level,
       "these issues" being both software issues and data --
   A.  Data.
   Q.  -- more generally data issues.  You are not
       suggesting -- actually let me break it down so we can be
       completely clear about what we're talking about.  You've
       got -- it is not a venn diagram, in fact it is the
       opposite of a venn diagram.  We've got a category of
       what I would call -- let's call them coding issues so we
       know exactly what we're talking about, and then we have
       a category of data issues.  Would I be right in thinking
       that the data issues that you encountered were far more
       frequent than the coding issues, the coding issues
       actually were quite rare?
   A.  We would always encounter data issues.  Those data
       issues could be caused by a coding issue, or they could
       be caused by something else and I would say they were
       usually caused by something else.
   Q.  I'm grateful.  So data issues that were the result of
       coding issues represented a very small proportion of the
       total world of data issues, yes?
   A.  Yes.
   Q.  I'm grateful.  And then of that -- we then look at data
       issues and we have data issues that result in
       discrepancies at branch level.
   A.  Yes.
   Q.  Would I be right in thinking that the number of data
       issues that cause discrepancies at branch level were
       very small compared with the total number of data issues
       with which you were wrestling?
   A.  My recollection is that the data issues we were dealing
       with were generally -- that it was quite a high
       percentage of -- they were causing problems at the
       branches.  I will ...
   Q.  You say causing problems at branches, there are all
       sorts of problems that would be caused at a branch.  For
       example you could have a configuration problem which
       means the machine isn't working properly.  That's a data
       issue, isn't it?
   A.  Yes, yes.
   Q.  That would be a problem at the branch but it wouldn't be
       a problem that would cause a discrepancy, would it?
   A.  So what I mean is from what I remember, a lot of the
       data issues that came in, we got them because there was
       a problem with the accounts, a balance mismatch
       somewhere.  This is how I remember it.  And if we could
       find a problem -- if you could pinpoint it to the data
       itself then you could perhaps find the problem and fix
       it.  If you couldn't find where the problem was in the
       data, or you couldn't understand why it had happened
       then there was the potential for it to be a software
       problem, so then you would pass it over.  It didn't mean
       there was a software problem, it meant there was
       a potential for it.
   Q.  Well, I'm grateful for that and actually I think you may
       have just agreed with what I'm about to put to you.
       Could I ask you to go to Mr Parker's first statement
       please.  It is at {E2/11/11}.  This is where Mr Parker
       says this:
           "Classifying an incident as a 'potential' software
       error does not necessarily mean that there was
       a software error ..."
           I believe that's the point you just made to me,
       isn't it?
   A.  Mm-hm.
   Q.  "... and, even if there was, it does not mean that the
       error was one that could have caused a financial impact
       in a branch's accounts ..."
           Would you agree with that?
   A.  Yes.
   Q.  And then he says in brackets:
           "... (a large proportion of these would be errors in
       numerous data centre resident systems that the
       subpostmaster never sees - errors were often as trivial
       as the use of 'Kg' instead of 'kg' on receipts)."
           Would you agree with that?
   A.  I can't remember.
   Q.  Well, I do suggest to you, Mr Roll, that even when you
       had a coding error that created a data problem in
       a branch, the number of data errors that actually
       created a branch shortfall was a vanishingly small
   A.  My recollection is that it was a big part of our job, so
       for me working there it didn't seem like a small
       proportion, a vanishingly small proportion.  Also
       just -- if there was a potential software issue and we
       passed it over for investigation, if they couldn't find
       a problem it didn't necessarily mean there wasn't
       a problem, it just meant they couldn't find it.  So ...
       I can't agree fully with what you have just said.
   Q.  Well, let me continue with what Mr Parker says:
           "As stated in paragraph 16 above, such errors were
       extremely rare."
           That's coding errors which caused a financial impact
       on branch accounts.  As a proportion of the work coming
       into the SSC, coding errors causing financial impact on
       branch accounts was extremely rare, would you agree?
   A.  Yes.
   Q.  Thank you.  And you say that problems would come in
       because there were discrepancies in branch accounts and
       that's how you would often see a problem, but of course
       there's a large number of possible causes --
   A.  Yes.
   Q.  -- for those problems, including user error, would you
   A.  Yes.
   Q.  And would you agree that a significant proportion,
       a large proportion of even the problems that got through
       to the SSC, to third line, they were due -- probably due
       to user error?
   A.  My recollection is that maybe 50% -- I'm trying to
       remember now.  I can't remember it fully.  But my
       recollection is that it was a pretty balanced mix as to
       either user error or maybe data corruption.
   Q.  So let me see if I understand your answer.  So what you
       are suggesting is of the problems that came in that were
       concerned with branch account discrepancies, you say
       your recollection is 50% you would have put down to user
   A.  Yes, I'm -- I'm trying to just think of it now and maybe
       50% was user error.  Certainly -- I wouldn't have liked
       to have said that the majority, or a vast majority was
       user error, but it is possible.
   Q.  Then let's move on to paragraph 42:
           "Mr Roll's suggestion in paragraph 10 that software
       issues in Horizon 'routinely' caused discrepancies in
       branch accounts is misleading.  In the vast majority of
       cases such an occurrence would cause a receipts and
       payments ... mismatch that would be flagged by the
       branch system as part of the balancing process ..."
           Is that right?
   A.  There were instances -- no, I can't -- I think -- I'm
       not sure.  I think that's right.
   Q.  "In the vast majority of cases such an occurrence would
       cause a receipts and payments ... mismatch that would be
       flagged ..."
           That's true, isn't it?
   A.  In the vast majority, yes.
   Q.  Then he adds:
           " ... (the Horizon system carries out
       self-consistency checks which generate alerts in the
       event of a receipts and payments mismatch that are
       picked up by SMC and incidents raised for the SSC) ..."
           Do you remember that?
   A.  In the vast majority of cases, yes.
   Q.  And it appears on MSU reporting.  Do you remember what
       MSU reporting is?
   A.  No.
   Q.  "These would then be investigated and resolved by the
           Is that right?
   A.  I can't remember.
   Q.  And then in 42.1 he says:
           "Since the introduction of Horizon in 1999 there
       have been 735 live incidents which refer to 'payments
       and receipt mismatch' (ie incidents recorded against
       components of the system providing Horizon service to
       Post Office rather than incidents raised against test
           Stopping there, would I take it that you are not in
       a position to challenge that figure?
   A.  No.
   Q.  "This figure has been obtained using a textual search
       across all incidents where the title or one of the
       incident updates contains all of the words 'receipts',
       'payments', 'mismatch'."
           Would you agree that there would be a fair way of
       trying to find the relevant PEAKs in this respect?
   A.  If you're looking for receipts and payments or mismatch
       then yes.
   Q.  I'm grateful:
           "It should be noted that this is not 735 unique
       incidents; there will be a lot of duplicates with the
       same root cause."
           Would you agree with that?
   A.  I can't disagree with it.
   Q.  "The only way to determine how many unique incidents
       there were would be to manually review all of the
           I take it you would agree with that?
   A.  Yes.
   Q.  "All of them were resolved."
           Do you accept that?
   A.  I can't comment.
   MR JUSTICE FRASER:  I don't think he would be in a position
       to -- I mean a question would be I suppose: from your
       time there do you remember any that were not resolved?
   MR DE GARR ROBINSON:  Yes, thank you.
   A.  I -- only vague memories of some things that may have
       not been -- we couldn't pin anything down specifically
       to code, or we couldn't find a problem, so it was
       assumed that the postmaster must have made a mistake.
   Q.  Well, you say "assumed", Mr Roll, wouldn't it be fair to
       say that when you're looking at -- we're now looking at
       software coding issues, you're looking for coding issues
       that may have caused a problem, would you agree that an
       enormously thorough amount of analysis is done to try
       and figure out how a coding problem may have caused this
       particular symptom; would you accept that?
   A.  Yes.
   Q.  And the SSC would keep going, people like the five
       specialists and indeed you when you were involved on
       those occasions, you would keep going until you had
       satisfied yourself, as far as any human being could,
       that it wasn't down to a coding error; would you accept
   A.  Sometimes there were times -- well, sometimes there were
       time constraints from what I remember, so you had a few
       days to find a problem and produce an answer.
   Q.  "Sometimes".  How often did that occur?
   A.  I couldn't say, but that's one of the things that sticks
       in my mind that there were occasions.  Very
   Q.  I'm grateful.  Let's move on to page 12 please
           "These incidents ..."
           We are now talking about receipts and payments
           "These incidents are reported as a result of
       self-consistency checks carried out by Horizon.  It
       should be noted that a R&P mismatch is not only caused
       by a software error.  It can also be caused by incorrect
       product reference ... data."
           That's right, isn't it?
   A.  Yes.
   Q.  Then 42.3:
           "Receipts and payments mismatches happened more
       often during the early life of Horizon ..."
           And there is a reference to a tab which I will go to
       in a moment:
           "My analysis of that data shows that there were
       around 8.6 such incidents per month on average between
       1 January 2001 and 31 December 2014 (417 out of a total
       of 27,005 incidents into SSC, or 1.5% of SSC incidents
       during that period)."
           We can go to the table if you like but I apprehend
       that you would probably give me the same answer that
       you're going to give me now which is that you are not in
       a position to challenge that analysis?
   A.  I'm not.
   Q.  So what we're talking about, if this analysis is right,
       is 1.5% of incidents going to third level, going to the
       third line which consist of payments and receipts
       mismatches and that means that less than 1.5% of
       incidents going into SSC would have been discrepancies
       in branch accounts caused by software issues, by coding
       issues.  Do you accept --
   A.  Yes.
   Q.  -- the logic of what Mr Parker is saying?
   A.  I don't know if the receipts payments mismatch would
       include other balancing problems or not.
   Q.  But I think you have already agreed that in the vast
       majority of cases such an occurrence would cause
       a receipts and payments mismatch?
   A.  I believe so.
   Q.  I'm grateful.  On the basis of that evidence -- would
       you give me a moment, Mr Roll, I do apologise.  One of
       my questions is being corrected by a junior who is far
       too able for his own good.
           At page 172 I think I did what is conventionally
       called an American misspeaking.  My question was "1.5%
       of incidents going to third level which consist of
       payments and receipts mismatches and that means less
       than 1.5% of incidents going into SSC would have been
       discrepancies in branch accounts caused by software
       issues" and I'm corrected, I should have said:
       potentially caused by software issues.
   MR JUSTICE FRASER:  I think you meant potentially caused by
       coding issues.
   MR DE GARR ROBINSON:  I really do mean that.
   MR JUSTICE FRASER:  But you then did go on to say "by coding
       issues" after software issues.
   MR DE GARR ROBINSON:  But the key word is "potential".
       I wasn't seeking to suggest to you -- I was seeking to
       suggest to you there is a wider category.  You recognise
       of course there are potential coding issues and then
       there are actual coding issues and the former is
       a larger class than the latter.  We have lots of venn
       diagrams today and what I was suggesting to you there
       was I was talking about the larger class.
   A.  Yes.
   Q.  That is what must be less than 1.5% --
   A.  Yes.
   Q.  -- would you accept that logic?
   A.  Yes.
   Q.  Thank you, Mr Roll.
           Now, if you encountered a software error causing
       a branch -- well, actually, I'm so sorry, let's go back
       to paragraph 19 of your first statement which is at
       page 3 {E1/7/3}.
           Could I suggest a way of rewriting paragraph 19 that
       would make it if not more accurate, at least read more
       fairly.  You start by saying:
           "In summary, the issues with coding in the Horizon
       system were extensive."
           But it would be fair to say "but represented a tiny
       proportion of the work coming into the SSC", would you
       accept that?
   A.  Sorry, which one are we looking at now?
   Q.  I'm so sorry, paragraph 19 on page 3 of your first
       witness statement.
   MR JUSTICE FRASER:  I don't think the witness has reached it
       yet.  He is using the hard copy.
           Look under tab 7 and then look at the penultimate
       page of your witness statement which should have
       {E1/7/3} in the bottom right-hand corner.  Do you see
   A.  Yes, thank you, my Lord.
   MR JUSTICE FRASER:  It has paragraph 19.
   MR DE GARR ROBINSON:  Mr Roll, I should have done that and
       I do apologise for not having seen your predicament.
           You start by saying:
           "In summary, the issues with coding in the Horizon
       system were extensive."
           Would it be fairer to say that the issues with
       coding in the Horizon system represented a very, very
       small proportion of the issues that the SSC dealt with?
   A.  From the evidence you have shown me just then I would
       agree with that.
   Q.  And then you say:
           "Furthermore, the coding issues impacted on
       transaction data and caused financial discrepancies on
       the Horizon system at branch level."
           Would you accept that the proportion of coding
       issues that had that effect was even smaller?
   A.  Yes.
   Q.  And the next sentence:
           "It was those issues that I, and other colleagues at
       Fujitsu, were routinely working on daily."
           Well, that's not really right, is it, Mr Roll?  You
       and others were not working on these issues daily, were
       you, they were a very small proportion of the work that
       was being done?
   A.  From the bugs then yes.
   Q.  Thank you.
   MR JUSTICE FRASER:  By bugs you mean coding?
   MR JUSTICE FRASER:  I seem to be the only person who cares
       about what words are used.
   MR DE GARR ROBINSON:  So if we could go back to paragraph 10
       please on page 2 of the same document {E1/7/2}.  You say
       that -- you refer to software issues, which we now know
       is a wider category, it's both coding issues and data
           "... could and did cause financial discrepancies ...
       if we were unable to find the cause of the discrepancy
       then this was reported up the chain and it was assumed
       that the postmaster was to blame."
           When you say "reported up the chain", to whom was it
   A.  To our manager and that's as far as I knew.  I don't
       know where it went after that.
   Q.  And why would you report it up the chain?  Why wouldn't
       you just complete the PEAK in the normal way?  There
       seems to be some significance attached to a process of
   A.  Usually, from what I remember, there was some pressure
       on us to come to a resolution, a solution.
   Q.  Oh, you mean -- I see, so do I understand this, that it
       wasn't -- the management didn't want you to close your
       eyes to issues of that sort, management wanted you to
       figure out what the problem was?
   A.  Quickly.
   Q.  That was the encouragement that --
   A.  They wanted a quick answer, yes.
   Q.  No, let's take this in stages.  They wanted you to
       figure out what the problem was, is that right?
   A.  Yes.
   Q.  And now you have added the word "quick".  Why do you say
       quick --
   A.  "Quick" was perhaps the wrong word.
   Q.  -- do you think that's entirely fair?
   A.  As I recall, sometimes we felt under pressure to
       provide -- to examine the data and give an answer within
       the timescale.
   Q.  So sometimes the SSC was busier than other times and
       when the SSC was busy then sometimes you would say,
       what, you didn't have as much time as you would have
       liked to have got to the bottom of the problem?
   A.  Yes.
   Q.  And in that situation where you hadn't done enough you
       would report that situation up to your manager, would
   A.  Yes.
   Q.  And that would be Mr Peach, would it?
   A.  Yes.
   Q.  And what would Mr Peach say?
   A.  I don't know what he said to his -- to the management
       above that, I've got no idea.
   Q.  But it would be fair to say -- I think I'm -- I hope I'm
       drawing the right inference from what you said a minute
       ago -- it would be fair to say that Mr Peach would want
       you to bottom out the issue, to get to the bottom --
   A.  Yes.
   Q.  -- of that issue that was reported to you.  He didn't
       want you to close your eyes to a problem, did he?
   A.  No.
   Q.  I'm very grateful, Mr Roll.
           When you encountered a coding issue which had --
       well, did you ever encounter a coding issue which had
       a financial impact on the branch account?
   A.  I can't remember.  There were times when I suspected
       something like that and passed it on, from what I do
       remember, to one of the other --
   Q.  To seek more senior people --
   A.  Yes.
   Q.  -- who might be able to look at it in more greater
   A.  Yes.
   Q.  But you don't recall ever having encountered, in all the
       PEAKs that you worked on, a coding issue that definitely
       caused a financial impact?
   A.  I don't recall discovering one, no.
   Q.  I'm grateful.
           So my next question was going to be when you were in
       that position, what would you do, but I apprehend that
       you won't be able to answer that question?
   A.  I can't remember.
   Q.  I was going to ask you about identifying all the
       branches that were affected by that coding issue but
       I apprehend that it's just not something you can speak
       to, is that right?
   A.  No.
   Q.  I see.  Can you just give me a moment.  I'm sorry to
       keep you waiting.
           So if we could go to {E2/11/3}.  Paragraph 17,
       Mr Parker says:
           "On the very rare occasion that a software problem
       which could cause a financial impact in branch accounts
       arose, it would be investigated and resolved and Fujitsu
       would determine its impact on the Horizon estate and
       inform Post Office of any financial impact on branches
       so that they could be resolved."
           Would I be right in thinking that's just not
       something that's within your experience because you were
       never in that situation?
   A.  Yes.
   Q.  Then I don't need to go there.
           My Lord, it occurs to me that now might be
       a convenient moment.
   MR JUSTICE FRASER:  All right.
           Mr Roll, I said at the break you were going to go
       overnight, so if you could come back tomorrow please,
   A.  Yes.
   MR JUSTICE FRASER:  And remember what I said about not
       talking to anyone about your evidence.  That doesn't
       mean that you can't talk to anyone about anything, but
       you can't talk to anyone about your case or your
   MR JUSTICE FRASER:  Couple of housekeeping things, we may as
       well deal with them now.  I don't think I have had
       a file next of your witness statements for next week.
   MR DE GARR ROBINSON:  No, that's in hand, my Lord.
   MR JUSTICE FRASER:  I would also like a single -- and this
       can be done as a running file.  The PEAKs that are being
       used in cross-examination -- and I think there have only
       been two or three of them so far -- if I can just have
       a hard copy of whichever ones are in fact used and those
       just put in a file and that file can be updated each
   MR DE GARR ROBINSON:  My Lord, we can prepare that.  That
       would be helpful.
   MR JUSTICE FRASER:  At the moment I think there is just the
       ones used in opening plus the three from today, so it's
       not going to be very difficult.
   MR GREEN:  My Lord, would you like as we go along to have
       a second section with the associated KELs?
   MR JUSTICE FRASER:  Yes, if that's going to be useful.
   MR GREEN:  I think typically you may see both.
   MR JUSTICE FRASER:  All right.  So that's that.
           You were going to tell me how many documents had had
       to be disclosed result of your recollection, or the
       redaction review.
   MR DE GARR ROBINSON:  I was.  My Lord, may we do it tomorrow
   MR JUSTICE FRASER:  Do it tomorrow, yes, we will do it at
       the end of tomorrow.
           I would like both your openings in Word but not the
       appendices.  And then I thought I would just give you an
       outline on the current plan on the judgment in common
       issues.  I've had a list of typos from the claimants.
       That has not affected the intended timetable to hand it
       down on Friday.  I imagine I will have some -- I think
       the time I asked for them was 5 o'clock, so I know it is
       a different team.
   MR DE GARR ROBINSON:  I'm afraid I know nothing about --
   MR JUSTICE FRASER:  No, it is just on the basis that it is
       the Post Office here, I thought I would just say.
       Assuming they are in the normal scope of typographical
       errors, et cetera, I'm still on track to be able to hand
       it down finally on Friday at 12 o'clock, but if that
       changes I will tell you tomorrow.
           Then one grammatical point which you may or may not
       find amusing, I do, but I thought I would ask you.
       "Subpostmasters" in the draft judgment you will see
       I abbreviated to a three letter abbreviation, or TLA, of
       SPM.  I have put it as the indefinite article "an SPM".
       Mr Warwick's helpful corrections have suggested that
       throughout it should be changed to "a SPM".  I don't
       mind because it takes a single keystroke, but in terms
       of grammar I thought it should be "an" SPM but I thought
       I would explore your views before I hand down a 180,000
       word judgment riddled with a potential grammatical
       inaccuracy.  So Mr De Garr Robinson?
   MR DE GARR ROBINSON:  Your Lordship will have seen from our
       submissions we say "an" SPM.
   MR GREEN:  My Lord, I have no problem with "an" SPM at all.
   MR JUSTICE FRASER:  You hadn't spotted that.  I'm not sure
       I actually know with absolute precision what the rule is
       for abbreviations, but we will probably stay with "an".
           Anything else?
   MR GREEN:  No, my Lord.
   MR JUSTICE FRASER:  10.30 tomorrow.
           See you tomorrow, Mr Roll.  Thank you all very much.
   (4.30 pm)
        (The court adjourned until 10.30 am on Thursday,
                         14 March 2019)