Tuesday, June 4, 2019

Horizon trial day 14: It's all relative

Anthony de Garr Robinson (r)
We are in the hard grind of the Horizon trial. If you want to read the live tweets - you can find them collated here on a single, beautifully formatted webpage thanks to the threadreader app.

I've just had a look and they do come across like a coded cry for help, but they will take you far deeper into the issues than what follows. Which is necessarily impressionistic (not least because I don't have the transcript of today's proceedings yet).

UPDATE: yes I do.

Mr de Garr Robinson (QC for the Post Office) is an excellent barrister, and Jason Coyne (independent IT expert for the claimants) is a deeply knowledgeable computer wizz, but I am not convinced the august and ancient art of adversarial cross-examination is the correct framework around which to hang a helpful discussion about IT. It's like challenging an algorithm to a duel.

Mr de Garr Robinson spent the morning establishing something we knew at the beginning of the trial, which is that Mr Coyne believes the Post Office's Horizon IT system to be "relatively" robust. David Cavender QC, another one of Post Office's barristers used to half-joke in the first trial that the court appeared to be trying the Subpostmaster contract. Today it felt like we were trying an adverb. Or deep space.

Mr Coyne was not prepared to rate Horizon out of five. He was not even prepared to say whether Horizon was better or worse than other equivalent large retail/banking systems. What he did agree, is that compared to other large retail/banking systems Horizon is relatively robust. Nowadays, that is.

Mr Coyne felt that when Horizon was first launched there was a period when it was "less" robust than it is now, and that when Horizon Online was brought in Horizon again became "less" robust, and that trying to come up with an absolute assessment of its robustness at any given period in its 20 year existence would be an impossible task.

Mr Coyne was not trying to be oblique, and the exchanges between Mr de Garr Robinson and Mr Coyne were marked by politeness and respect - I did actually feel that both "sides" were doing their best to get down into the meaning of Mr Coyne's evidence in a way that would properly assist the judge in his decision-making - but if I were a claimant with no previous experience of this process and I'd been parachuted into court to witness what was said today, I'd be scratching my head, to say the least.

Mr de Garr Robinson's general thesis was that Mr Coyne had a downer on Horizon in his first report (dated 18 October 2018) and that when given more disclosure about Horizon, Mr Coyne was forced to accept in his second report (dated 1 Feb 2019) that Horizon was actually better than he thought it was. ie the more Mr Coyne found out about Horizon, the better he realised it was, yet, Mr de Garr Robinson suggested, he seemed strangely loathe to admit it.

Mr de Garr Robinson also did a good job of challenging some other adverbs used by Mr Coyne in his report. Mr Coyne's opinion that a cost-benefit analysis was "often" used when discussing whether or not to fix or manage a bug in the Horizon system took about an hour of our day. Mr Coyne accepted that "often" was perhaps the wrong term to use, especially after Mr de Garr Robinson demonstrated one of his examples was completely wrong and then challenged Mr Coyne to come up with more than a single other example in twenty years.

At this point the judge intervened and asked Mr Coyne to come back to court tomorrow with anything which would justify his assertion cost-benefit analysis is "often" or even "generally" used when making decisions about fixing bugs.

Whilst all this was wanging on, I was thinking about a Subpostmaster, stuck with responsibility for a discrepancy they felt was a clear error, and having to swallow it or flag it, or deal with it in a way which allowed them to continue running their business. What did Horizon's relative robustness have to do with them? The space shuttle programme was relatively robust. As is the Boeing 737.

As Mr de Garr Robinson said today, Horizon has processed around 49 billion transactions since its inception. Everyone on both sides of this litigation seems to accept as fact that Horizon is an ecosystem so vast it is capable of producing countless known and unknown errors (ie you can't count an error you don't know exists).

The real issue for me is whether, due to human or undetected computer error, a whole bunch of people have been held responsible for something which wasn't their fault.

And whilst Mr de Garr Robinson suggested that was exactly why we were there, I'm not sure, today at least, we got any closer to an answer.

Luckily, there are three more days of Mr Coyne's cross-examination to come.

If you would like to get across either of Mr Coyne's 200+ page reports into Horizon, they have been released by the court and are available to read here. I have requested the joint statements, which are documents agreed between Mr Coyne and the Post Office's independent IT expert, Dr Robert Worden. I hope to receive them shortly.

*******************
If you can, please help keep this crowdfunded public-interest journalism project going by chucking a few quid in the tip jar below. Contributors who give £20 or more will start receiving regular "secret" emails which have all the info and gossip about this litigation as it makes its way through the courts.

If you want to find out a little bit more about the underlying story, click here.

Choose an amount