Convicted Post Office workers have names cleared

>Does anyone know how the courts rendered these verdicts?

I am wondering too if they maybe had "experts" testify , and those experts were just clueless.

I don't know the full details of how the judges operated during the prosecutions, but one of the bizarre details is that the post office has the oldest prosecution and investigation team in the world, predating the police force, dating back to highwaymen robbing postal services on horseback. In other words, the post office executives seem to have made a policy decision around these financial discrepancies, the post office managers gathered evidence on their behalf under the guise of helping their employees, and the post office prosecution team prosecuted them in court. Very little unbiased external opinion was brought in. Even the police seemed to have decided early on that they didn't have jurisdiction in the post office for the above reasons. Its an extremely bizarre detail that, as a UK resident, I had no idea about until these recent revelations.

The judgment has a few details.

https://www.judiciary.uk/judgments/hamilton-others-v-post-of...

There's possibly something on Bailii. (A small volunteer run project, but the website is tricky to use.) Searching for [post office horizon] reveals a few cases.

Here's one useful transcript and discussion: https://journals.sas.ac.uk/deeslr/article/view/2217

Everyone except the victims is to blame here. The courts should have done better, but in the UK these people would have been unable to get a public defender and unable to afford their own defence either. So many would have plead guilty and others had to try and manage the defence themselves. The result is predictable.

A wider question is why no one asked why so many people with the same job were being charged and convicted/pleading guilty. Fraud is a rare crime and rarely prosecuted. So many Cases in one profession is fishy all by itself.

The law presumes that computer systems operate correctly unless there is evidence to the contrary.

This is a reasonably common thing - for example the law also assumes letters are delivered after being posted unless there is evidence to the contrary.

UK courts also sent a woman to prison for infanticide, basically off of the testimony of a single doctor who didn't know how statistics work:

https://en.wikipedia.org/wiki/Sally_Clark

I agree the issue is a legal system bug.

> Software is buggy. Since when does a court accept the uncorroborated report of software, without any other corroborating evidence, and sentence someone?

Former PwC Auditor here , the court accepts evidence from independent IT Auditing company who are qualified to perform an audit on the system and asses the reliability of the system or from the vendor himself who provide proofs that the system he sells is reliable ( testing , certifications etc..)

If you're developer you'll probably agree with me that this approach is of course a "non-sense" because no software was ever created "bug-free" or can be really defined as "reliable".

Yet , that's how the court works : "The vendor says the software has no bugs , thus the court is rejecting the objection of buggy software. The court found you guilty"

I'm sadden by this news because it depicts how much the mixture of "bad software" and "corporate/enterprise software" are tied together and how much it can impact people on their daily life with irreversible impact.

The answer to this question is that senior executives within the Post Office knew that Horizon had faults and deliberately withheld the information when prosecuting the sub-postmasters.

This is a fundamental breach of the prosecutor's duty to disclose any evidence that undermines the prosecution case or supports the defence case. This duty continues to exist after conviction so timing of knowledge is irrelevant.

The judgment is telling in that there are records in which Post Office officials made statements that minutes of meetings about faults in the Horizon system should not be taken so as to avoid having to disclose them in proceedings.

It is corporate failure on an unimaginable scale and three convicted individuals have died before having their convictions quashed.

Yesterday's judgment is long but very readable. I would anticipate further fallout and understand there may be a live police investigation on the basis that several individuals may have perverted the course of justice by either proceeding with prosecutions or omitting material evidence from testimony.

In one case (Seema Misra) the judge asked the jury this in his/her summing up:

“There is no direct evidence of her taking any money... She adamantly denies stealing. There is no CCTV evidence. There are no fingerprints or marked bank notes or anything of that kind. There is no evidence of her accumulating cash anywhere else or spending large sums of money or paying off debts, no evidence about her bank accounts at all. Nothing incriminating was found when her home was searched...Do you accept the prosecution case that there is ample evidence before you to establish that Horizon is a tried and tested system in use at thousands of post offices for several years, fundamentally robust and reliable?”

The jury pronounced Seema Misra guilty.

There are many more details in this very comprehensive report by Private Eye: https://www.private-eye.co.uk/special-reports/justice-lost-i...

Legal reform is like software refactoring. You always know it needs to happen. It rarely ever does. Even if it does, there's still bugs.

> Since when does a court accept the uncorroborated report of software, without any other corroborating evidence, and sentence someone?

Since before software has had bugs. Law is the human-governance form of software. It's just as shitty as software is, for the same exact reason.

"Bad software" is only part of the story.

From what has been reported in the media, the Post Office and the contractor knew that the software had issues but chose to cover it up and to go after innocent individuals instead. In some instances people were also made to sign settlements preventing them from criticising the software.

This is why this is such a big scandal in the UK. The software was buggy but people lied and sent innocent people to jail to cover it up.

This occurs for other things across a variety of spheres, but defence barristers are more or less experienced in some of those spheres, and lay people have more or less knowledge in some spheres.

If a police officer tells the jury that they are 100% sure the person they saw with the knife was the defendant, the defence understands how and to what extent they might be able to persuade the jury that the officer is mistaken about this, or how else to undermine this testimony, and they won't accept it if the officer seems to suggest that, since they're sure this was the person with the knife therefore that person stabbed the victim. There are lots of reasons why, even if the jury is convinced their client is lying and had a knife, that doesn't necessarily mean they stabbed the victim.

In contrast when an "IT expert" tells the jury that 100% the only way this database entry would appear is if the defendant stole money, the defence legal team may be unsure how to persuade a jury (since they aren't IT experts) that this might not be so, and they may end up allowing this to go unchallenged even though I would know to ask lots of questions about access, logging, test strategies, and so on.

Or if a medical doctor tells the jury that the only way this baby gets so many broken bones is that her parents were deliberately shaking her because she was crying, and so they're guilty of murder, the lawyer isn't a medic, and neither is the jury. So who is there to tell them that er, actually there can be other causes and the prosecution needs something else?

Light Blue Touchpaper covered some cases where it seems likely what happened is that crap implementations of EMV ("Chip and PIN") were defrauded by crooks who understood where the flaws were and how to exploit them, but the banks went after their customers, telling juries that the machine is infallible and if it says they used a PIN, that must be true, even though the researchers show various ways it might not in fact be true. Further, bank employees will cheerfully tell a jury that there's no way for other employees to discover a customer PIN, even though meanwhile the bank is firing employees it knows did just that. We ultimately cannot entirely insulate the justice system against people who lie under oath. We can only punish them when they get caught.

So, to the extent anything should be done here, you'd want to start with say, not rewarding the people at Post Office Ltd and in the Horizon team who lied about this and tried so hard to prevent justice being done. Two guesses whether the present government chose to do that...

Not only the software was buggy, the court was buggy as well ...

Apparently, some of the defendants were also so intimidated by the problem they tried to replace the "lost" money with their own money.

People believe machines because they don't lie. There is a cognitive dissonance between knowing that software is written by humans who make lots of errors and the believe that machines don't make errors.

Courts are led by humans. Judges (and juries) and in particular old judges (and older juries) often don't have a good grasp on technology and injustice like this happens. Luckily there is the appeal process in most modern countries so that such misdeeds can be corrected. Unfortunately a lot of people were hurt by this. Hopefully some good education of justice officials can come out of this.

There is a paper here [1] based a lot on these cases calling for a review of the presumption that computer evidence is correct.

It is an absolutely crazy situation where the onus is on the person disputing the computer evidence to show that something is wrong. Even in the middle ages they had a system in place where both parties would have reliable evidence to deal with non reputation in financial transactions. They used a split wooden tally sticks [2], yet now because it is on a computer, and computers are supposedly rarely wrong the courts will allow an organisation to produce something from essentially their own 1 sided ledger and it is up to the other side to prove it wrong. The sub postmasters had no evidence, no ledgers or paper receipts of their own, they were expected to rely entirely on this system to be correct, when it wasn't the entire justice system was rigged against them with the only possibility of winning was to prove that a multi million £ IT system was not working correctly. This would have been hard enough anyway, but the company who built it and the one using it were both willing to lie to cover up its failings.

[1]https://journals.sas.ac.uk/deeslr/article/view/5240

[2]https://en.m.wikipedia.org/wiki/Tally_stick

What are the chances we'll see jail time for any of the people responsible? I'm guessing zero?

I find it hard to believe that the approach taken - to prosecute sub post masters in this way - didn't make it to board level (indeed if it didn't then that by itself would be a corporate failure).

Individual managers may have had incentives that would lead to take this approach but the PO Board should have had an overview and have been able to correct this.

In any event a great scandal and very sad as you say.

It still exists! It's now mainly a mailing list, which I'm still subscribed to and still receiving. Unfortunately right this second the website seems to be down, but it's: http://www.risks.org

People getting forcibly removed from overbooked flights was another. When policies rely solely on computer automation, and human employees aren't allowed to make common sense ad hoc decisions, you end up with some ridiculous injustices (however petty).

If one were writing a novel, the thief would be hiding their tracks by being in a key place feeding the prosecution.

But really there is a track record of court systems latching onto criminal prosecutions via questionable technologies from handwriting analysis to lie detectors et al.

Yes, someone from Fujitsu who didn't deliver quality software, refused to accept it _might_ be buggy.

>In the same speech, he said that the Post Office would work with the government to compensate the employees who were affected by Horizon’s inaccuracies.

Fujitsu won't even be held financially accountable for the suicide, but tax payers will. To me, it sounds like a great deal, delivery shitty software, bounce back any questions about quality, bugs, refuse to re-test software for given scenarios for years, but still get paid.

> could not have happened

Could still happen. But would be less likely.

this is indeed the mind-boggling part for me. nothing was done to prove these people guilty but simply trust the output of a piece of software.

in the deep-mind world we live in today that's a scary thing to do.

i am going to have to agree with the group that feel unhappy with the monetary outcome.

As I understand it, the issue is there was never a second record.

Imagine you run a post office, people pay you cash to post their letter/parcel instead of putting a stamp on it.

You are meant to enter into the till every time you do that. Every pound you get this way, you owe 50p to RM who will pickup and deliver whatever is in the bag at the end of the day.

You diligently and honestly enter every pound you get into the till. But the till is broken, every time someone buys a Mars bar, it adds one to the count of objects posted.

At the end of the month, royal mail think you posted 1000s of extra items. They demand 1000s of extra 50ps. You didn't and don't count every item, that's the tills job. But somehow your losing money. You don't know how, RM don't care they just demand payment. Sooner or later you cannot pay (or refuse as the number must be wrong).

Then the courts get involved...

Not sure if you're being sarcastic, but yes, they may have had all those things, and still had this bug, because cargo cult testing is not usually effective at finding bugs.

I think in this case they simply mean it's the highest number of defendents who have succeeded in having their convictions overturned.

It's funny because when it was the little guy it deserves jail time, when they found it was a software company it's negligence and no jail time. That being said I understand why, but try convincing the people that were jailed of that.

Do you think that using your favorite cool language and stack there will never be logic errors? Could be worse, some idiot developer would be even less humble if he thinks he is using some bullet proof language.

> On the end who's accountable for this mess?

The judicial system. You don't lock someone up because some powerful/big company say they stole. My guess is they took the company for its own word.

The Post Office is responsible because they drove the process of:

1) buying broken software and not checking it was fit for purpose

2) Vigorously investigating post masters and not relenting even when many of the people being investigated claimed there must be a problem with the software

3) (and this is the kicker) continuing when they started to realise that maybe the software was the problem

Personally I think members of the board at the time need to face some kind of criminal, or at least civil, action here.

> Do those developers even realise it was their incompetence that caused untold misery?

Whilst the developers did ship broken software, I don't think you can lay the blame for the misery at their door directly - that should be on the management who, when presented with evidence that the software was garbage, doubled-down and decided to prosecute rather than running tests, etc.

The project managers who demand one-hour turnaround on every feature request and mandatory overtime are to blame.

The full headline feels misleading. It's not that no-one wanted to admit it could be wrong. No-one wanted to admit it was wrong. There were plenty of people involved who were well aware it was wrong, they just didn't want the consequences of that admission.

Basic professional ethics. Don't do or participate what you're not comfortable with. Hold your ground and let them either do things right or fire you.

This happened to me at a Big Name university where they wanted me to rush punching major holes in a secure private network for credit card processing to make it more convenient for a contractor to be able to remote in. I told my boss "it's not going to happen on this timeline without a security review" and was told to do it or resign. That was it.

I don't get this from the story. No one from the Post Office or Fujitsu has been held accountable. Literally no one and this case, these cases, were ongoing since 1999.

Yes, people also built the buggy software. But the role of software here can’t be overlooked. Because people trust software to run the modern world, there is a high level of trust that this software is working properly. The article says that this trust was misplaced in this case and it caused serious harm. It doesn’t just blame the existence of software, but questions the way in which it is built, used, and the ways in which society places trust in it.

> The bad software was just one part of the input to that decision.

To play devil's advocate was there evidence of malfeasance besides the software?

If not, it is a bit disingenuous to say that the software did not send them to jail. If it was a witness who was mistaken and that was the only evidence, in standard english, we would say the witness sent them to jail.

We've changed the URL now. The submitted URL (https://www.theverge.com/2021/4/23/22399721/uk-post-office-s...) had the bad bad software title.

?

The system miscalculated some figures and assumed subpostmasters (delivery people) were stealing money. There was likely no theft occurring.