Mr Bates vs The Post Office
Discussion
Mojooo said:
outnumbered said:
Beer says that they only realised this yesterday.
Sounds like something out of a Hollywood movie where the smoking gun is found the night beforeUnderhand tactics?
mikeiow said:
Any top sysadmin (particularly one defined as a "Distinguished Engineer") will know where the bodies lie - they will know that there will always be a possibility of a back door, and it is very, very hard to imagine Jenkins not knowing this was done.
yep, even for lowly IT support like I was in the 80s there was an override password for the gov systems we installed. it was......<drum roll>....... ( + ) = #
and you could change anything
mikeiow said:
I'm a little behind (had to pause things), but Jenkins refuses to call it tampering, where we absolutely know that numbers WERE changed and people WERE prosecuted and jailed off the back of that tampering.
What a nasty piece of work he is: genial demeanour or not, he is a bad person.
Beer appears to be on him.
I'm not sure we do actually know that, it's been suggested by various people.What a nasty piece of work he is: genial demeanour or not, he is a bad person.
Beer appears to be on him.
The legal issue causing the unsafe convictions isn't that numbers were changed and people got jailed, it's the fact that numbers COULD theoretically have been changed, but that wasn't disclosed.
siremoon said:
simon_harris said:
Every developer I have ever worked with has had 1000% confidence that the system they built was correct and working as intended.
We used to have an in house developed data processing solution that ran like dogs
t in a virtual environment and the dev team were adamant we were not allocating enough RAM, we proved conclusively they way they had written it mean that it used almost no RAM but did need a big frontside bus so as not to overwhelm the available CPU.
we built a dedicated physical machine to run it and ran it about 5000% better than compared their required/advised virtual spec.
Every developer I have ever worked with has had 1000% confidence that the system they built was compliant with the spec they were given.We used to have an in house developed data processing solution that ran like dogs
t in a virtual environment and the dev team were adamant we were not allocating enough RAM, we proved conclusively they way they had written it mean that it used almost no RAM but did need a big frontside bus so as not to overwhelm the available CPU.we built a dedicated physical machine to run it and ran it about 5000% better than compared their required/advised virtual spec.
I've rarely seen a spec that represents what the customer actually wants.
By the time I get to test a feature, I'll have the original spec, what did the original spec say should happen, then what testing was done around the spec to ensure it didn't fail when not played in the correct manner? Basic flows alone are never enough, you need some alternate ones and give the tester time to rip it apart every which way and document how.
It's made me be a bit more thorough on my comments and release notes even in my little world seeing a now elderly man have to remember wtf he did 20 years ago on just another project (guilty or not guilty). although I'm thankful that I can openly question things and my superiors take action and notice, which probably makes it a totally different atmosphere to Fujitsu at the time
simon_harris said:
You have to respect Sir Wyn - he is absolutely sharp as a tack and times his moments perfectly.
Jenkins " I wouldn't have written it like that"
Sir Wyn " But you didn't write it did you?"
"and you didn't write that either"Jenkins " I wouldn't have written it like that"
Sir Wyn " But you didn't write it did you?"
Must admit I'm getting a bit lost on some of this now.
LimmerickLad said:
"and you didn't write that either"
Must admit I'm getting a bit lost on some of this now.
I think the gist of it is that they're picking at all the places where he's not acted properly as an expert witness. Must admit I'm getting a bit lost on some of this now.
Personally I don't see him as having acted maliciously, there seem to be examples where he pushed back on things he was asked to do that he was unhappy about, although probably not enough. But also this wasn't really his main job, or anything that he was particularly interested in, he was a senior technical architect/designer with his own project work to do, so I can easily see how at some point he'd just give up arguing and go with whatever the PO lawyers were happy with.
So it looks to me more like he didn't really appreciate what was required, nonetheless tried to make a decent go of doing what he was asked, but fundamentally this wasn't to the required legal standards, and unfortunately the PO then used that as a weapon against the SPMRs.
outnumbered said:
LimmerickLad said:
"and you didn't write that either"
Must admit I'm getting a bit lost on some of this now.
I think the gist of it is that they're picking at all the places where he's not acted properly as an expert witness. Must admit I'm getting a bit lost on some of this now.
Personally I don't see him as having acted maliciously, there seem to be examples where he pushed back on things he was asked to do that he was unhappy about, although probably not enough. But also this wasn't really his main job, or anything that he was particularly interested in, he was a senior technical architect/designer with his own project work to do, so I can easily see how at some point he'd just give up arguing and go with whatever the PO lawyers were happy with.
So it looks to me more like he didn't really appreciate what was required, nonetheless tried to make a decent go of doing what he was asked, but fundamentally this wasn't to the required legal standards, and unfortunately the PO then used that as a weapon against the SPMRs.
outnumbered said:
Jenkins in a bit of trouble this morning.
The inquiry have belatedly provided him with evidence that he was sent a document that set out the role of an expert witness, along with some agenda information for a meeting he was going to attend. Beer says that they only realised this yesterday.
Hence some of the statements Jenkins made yesterday are being revisited. I suppose he's on slightly stickier ground as a result, because he now can't say "nobody told me", it's more "yes, someone did send me this thing in passing once but I didn't take as much notice of it as I apparently should have done"
I'm just catching up. In fairness to Jenkins, reading that email to him of 05/06/2006, it isn't the case (as Beer said) that the letter was sent "directly" to GJ. In fact it wasn't; it was sent "directly" to others, one of whom forwarded it to GJ with a covering note that (if it were me receiving it, at 19:43 UTC (!!), with the instruction that I might be needed to contribute *tomorrow*) would lead me to focus on the thing I could contribute to.The inquiry have belatedly provided him with evidence that he was sent a document that set out the role of an expert witness, along with some agenda information for a meeting he was going to attend. Beer says that they only realised this yesterday.
Hence some of the statements Jenkins made yesterday are being revisited. I suppose he's on slightly stickier ground as a result, because he now can't say "nobody told me", it's more "yes, someone did send me this thing in passing once but I didn't take as much notice of it as I apparently should have done"
I really don't see this as a smoking gun for GJ. In fact, it is pretty extraordinary - and poor - to be emailing GJ about a "tomorrow" issue that late in the evening.
I really don't see how this is supposed to reflect badly on GJ. Furthermore, if this was the only time he received any sort of "briefing" as to how to operate as an expert, I really don't see what he's done wrong *in that regard.*
Worse still, the letter is addressed to Fujitsu, and uses the words "you" to refer to Fujitsu, not GJ. The document was forwarded to him, but was he supposed to take from this that he was the "you" referred to? Fujitsu would have such a duty of care, and would need to ensure its employees did likewise. Just forwarding on that letter did not meet any sort of threshold of Fujitsu complying with their obligations, briefing / managing GJ, etc.
The letter refers to all sorts of further attachments. Were those sent to GJ? This looks like a scan of a letter. If the attachments weren't provided to GJ, he would reasonably assume he wasn't meant to do anything with it.
I think Beer is honestly being rather unfair to GJ here, and that Fujitsu (if they're honestly trying to rely upon this one forwarded letter) are culpable.
But it is early in they day, and I'm sure more will come out.
LimmerickLad said:
outnumbered said:
LimmerickLad said:
"and you didn't write that either"
Must admit I'm getting a bit lost on some of this now.
I think the gist of it is that they're picking at all the places where he's not acted properly as an expert witness. Must admit I'm getting a bit lost on some of this now.
Personally I don't see him as having acted maliciously, there seem to be examples where he pushed back on things he was asked to do that he was unhappy about, although probably not enough. But also this wasn't really his main job, or anything that he was particularly interested in, he was a senior technical architect/designer with his own project work to do, so I can easily see how at some point he'd just give up arguing and go with whatever the PO lawyers were happy with.
So it looks to me more like he didn't really appreciate what was required, nonetheless tried to make a decent go of doing what he was asked, but fundamentally this wasn't to the required legal standards, and unfortunately the PO then used that as a weapon against the SPMRs.
Not as outrightly malicious but still an element of such.
skwdenyer said:
outnumbered said:
Jenkins in a bit of trouble this morning.
The inquiry have belatedly provided him with evidence that he was sent a document that set out the role of an expert witness, along with some agenda information for a meeting he was going to attend. Beer says that they only realised this yesterday.
Hence some of the statements Jenkins made yesterday are being revisited. I suppose he's on slightly stickier ground as a result, because he now can't say "nobody told me", it's more "yes, someone did send me this thing in passing once but I didn't take as much notice of it as I apparently should have done"
I'm just catching up. In fairness to Jenkins, reading that email to him of 05/06/2006, it isn't the case (as Beer said) that the letter was sent "directly" to GJ. In fact it wasn't; it was sent "directly" to others, one of whom forwarded it to GJ with a covering note that (if it were me receiving it, at 19:43 UTC (!!), with the instruction that I might be needed to contribute *tomorrow*) would lead me to focus on the thing I could contribute to.The inquiry have belatedly provided him with evidence that he was sent a document that set out the role of an expert witness, along with some agenda information for a meeting he was going to attend. Beer says that they only realised this yesterday.
Hence some of the statements Jenkins made yesterday are being revisited. I suppose he's on slightly stickier ground as a result, because he now can't say "nobody told me", it's more "yes, someone did send me this thing in passing once but I didn't take as much notice of it as I apparently should have done"
I really don't see this as a smoking gun for GJ. In fact, it is pretty extraordinary - and poor - to be emailing GJ about a "tomorrow" issue that late in the evening.
I really don't see how this is supposed to reflect badly on GJ. Furthermore, if this was the only time he received any sort of "briefing" as to how to operate as an expert, I really don't see what he's done wrong *in that regard.*
Worse still, the letter is addressed to Fujitsu, and uses the words "you" to refer to Fujitsu, not GJ. The document was forwarded to him, but was he supposed to take from this that he was the "you" referred to? Fujitsu would have such a duty of care, and would need to ensure its employees did likewise. Just forwarding on that letter did not meet any sort of threshold of Fujitsu complying with their obligations, briefing / managing GJ, etc.
The letter refers to all sorts of further attachments. Were those sent to GJ? This looks like a scan of a letter. If the attachments weren't provided to GJ, he would reasonably assume he wasn't meant to do anything with it.
I think Beer is honestly being rather unfair to GJ here, and that Fujitsu (if they're honestly trying to rely upon this one forwarded letter) are culpable.
But it is early in they day, and I'm sure more will come out.
Several years ago I worked for a UK company that had a US parent that had self-reported to the SEC for an FCPA breach. I was interviewed by a team of New York lawyers who had 3" thick print outs of emails and minutes going back a few years. They repeatedly expected me to explain why someone else had said something in an email several years before. They'd done a forensic search for specific keywords and all sorts of irrelevant stuff was being triggered. I'm sure I said "I can't remember" and "I wasn't aware" (and meant it) several times.......
It transpired none of us had been trained on FCPA when the Yanks had taken over so were pretty much in blissful ignorance of what we were now being held to (and not previously by UK law). The episode is now a poster boy case study for the SEC themselves on what companies should be doing as we were the only US company ever to be levied no fines by the SEC despite our admission of breaches. Several other big names got done for $Billions for similar but were found out rather than self reported.
I've just been pausing the video and reading the draft witness statement in the Castleton case.
The implications are breathtaking. Have a read of the paragraphs leading up to para 27. I'll type out a bit here, which relates to daily and weekly declarations of cash and stock. Users were expected to *manually* use an ID number every day to denote a cash drawer and, if they typed the wrong number in (intending it to be common), they'd get a larger underage/overage.
Any sensible UX person would have a configuration system that allowed the SPM to specify the tills, their IDs, the basis of counting (all drawers together, or declared separately), and so on, and then *prompts* to count "till X", not "please enter the till ID."
If this is indicative of the way in which Horizon was designed, I'm pretty stunned TBH.
The other thing that's clear is GJ wasn't architect of the whole process; he was the Horizon guy. He talks in later paras about things like cheques, and having no idea what the process would be if there was a mis-match. He seems to have been the most narrow of "architects," designing a system that seemingly operated in some sort of vacuum.
Staggering.
The implications are breathtaking. Have a read of the paragraphs leading up to para 27. I'll type out a bit here, which relates to daily and weekly declarations of cash and stock. Users were expected to *manually* use an ID number every day to denote a cash drawer and, if they typed the wrong number in (intending it to be common), they'd get a larger underage/overage.
Draft Witness Statement said:
As part of the Balancing process a similar cash Declaration is done using the Declaration IDs exactly like the ONCH declarations. In 2004, these were separate independent declarations, but in 2005 they were merged to use exactly the same process for both ONCH and Balancing declarations. As part of balancing all Declarations with separate IDs are taken and added up and then compared with the System Generated Cash figure which is produced by analysing all transactions within the cash account period. Any difference is highlighted to the subpostmaster as a <<Discrepancy>> which if the[y] accept, will adjust the system cash figure to match the amount of cash that has been declared by the subpostmaster. Therefore if he has <<doubled declared his cash>> and accepts the resulting discrepancy, the system cash figure will be adjusted to reflect this. I suspect that this may be what has happened. However it is still Mr Castleton's responsibility to spot this and not to accept the (false) discrepancy when he is balancing the system.
Wow. Please tell me the likes of GJ weren't in charge of what we'd now call UX (User Experience), and that genuine HCI [Human Computer Interaction] people were involved in what was then Europe's largest IT project. To rely upon SPMs and their (low wage) staff to accurately type in a till ID every day by hand and, if they get it wrong, to double-count all the cash in the building, is just insane.Any sensible UX person would have a configuration system that allowed the SPM to specify the tills, their IDs, the basis of counting (all drawers together, or declared separately), and so on, and then *prompts* to count "till X", not "please enter the till ID."
If this is indicative of the way in which Horizon was designed, I'm pretty stunned TBH.
The other thing that's clear is GJ wasn't architect of the whole process; he was the Horizon guy. He talks in later paras about things like cheques, and having no idea what the process would be if there was a mis-match. He seems to have been the most narrow of "architects," designing a system that seemingly operated in some sort of vacuum.
Staggering.
Mercdriver said:
Quote “ Staggering?
No it’s not it’s incompetent
I would agree with that. Looks like the Horizon system was not fit for purpose and because POL had battered down on price FUJ were not overly concerned about its BEDs.No it’s not it’s incompetent
Listening to GJ, it sounds to me that he was never fully briefed by both POL and FUJ as to his role as an Expert Witness and rather than have a dedicated person looking at all the legal cases, GJ became to go to guy although he was also doing other things so couldn’t be fully bothered and didn’t give it all the attention it should have done.
Whilst he is culpable to a degree, others are far more to blame both in POL ad FUJ.
skwdenyer said:
I've just been pausing the video and reading the draft witness statement in the Castleton case.
The implications are breathtaking. Have a read of the paragraphs leading up to para 27. I'll type out a bit here, which relates to daily and weekly declarations of cash and stock. Users were expected to *manually* use an ID number every day to denote a cash drawer and, if they typed the wrong number in (intending it to be common), they'd get a larger underage/overage.
Any sensible UX person would have a configuration system that allowed the SPM to specify the tills, their IDs, the basis of counting (all drawers together, or declared separately), and so on, and then *prompts* to count "till X", not "please enter the till ID."
If this is indicative of the way in which Horizon was designed, I'm pretty stunned TBH.
The other thing that's clear is GJ wasn't architect of the whole process; he was the Horizon guy. He talks in later paras about things like cheques, and having no idea what the process would be if there was a mis-match. He seems to have been the most narrow of "architects," designing a system that seemingly operated in some sort of vacuum.
Staggering.
Could you give a reference to that document you quoted ? I.e. where is it in the video, or on the website ? It's interesting if it really gives an obvious mechanism for the discrepancies (poor UI design leading to mistakes by a clerk).The implications are breathtaking. Have a read of the paragraphs leading up to para 27. I'll type out a bit here, which relates to daily and weekly declarations of cash and stock. Users were expected to *manually* use an ID number every day to denote a cash drawer and, if they typed the wrong number in (intending it to be common), they'd get a larger underage/overage.
Draft Witness Statement said:
As part of the Balancing process a similar cash Declaration is done using the Declaration IDs exactly like the ONCH declarations. In 2004, these were separate independent declarations, but in 2005 they were merged to use exactly the same process for both ONCH and Balancing declarations. As part of balancing all Declarations with separate IDs are taken and added up and then compared with the System Generated Cash figure which is produced by analysing all transactions within the cash account period. Any difference is highlighted to the subpostmaster as a <<Discrepancy>> which if the[y] accept, will adjust the system cash figure to match the amount of cash that has been declared by the subpostmaster. Therefore if he has <<doubled declared his cash>> and accepts the resulting discrepancy, the system cash figure will be adjusted to reflect this. I suspect that this may be what has happened. However it is still Mr Castleton's responsibility to spot this and not to accept the (false) discrepancy when he is balancing the system.
Wow. Please tell me the likes of GJ weren't in charge of what we'd now call UX (User Experience), and that genuine HCI [Human Computer Interaction] people were involved in what was then Europe's largest IT project. To rely upon SPMs and their (low wage) staff to accurately type in a till ID every day by hand and, if they get it wrong, to double-count all the cash in the building, is just insane.Any sensible UX person would have a configuration system that allowed the SPM to specify the tills, their IDs, the basis of counting (all drawers together, or declared separately), and so on, and then *prompts* to count "till X", not "please enter the till ID."
If this is indicative of the way in which Horizon was designed, I'm pretty stunned TBH.
The other thing that's clear is GJ wasn't architect of the whole process; he was the Horizon guy. He talks in later paras about things like cheques, and having no idea what the process would be if there was a mis-match. He seems to have been the most narrow of "architects," designing a system that seemingly operated in some sort of vacuum.
Staggering.
outnumbered said:
skwdenyer said:
I've just been pausing the video and reading the draft witness statement in the Castleton case.
The implications are breathtaking. Have a read of the paragraphs leading up to para 27. I'll type out a bit here, which relates to daily and weekly declarations of cash and stock. Users were expected to *manually* use an ID number every day to denote a cash drawer and, if they typed the wrong number in (intending it to be common), they'd get a larger underage/overage.
Any sensible UX person would have a configuration system that allowed the SPM to specify the tills, their IDs, the basis of counting (all drawers together, or declared separately), and so on, and then *prompts* to count "till X", not "please enter the till ID."
If this is indicative of the way in which Horizon was designed, I'm pretty stunned TBH.
The other thing that's clear is GJ wasn't architect of the whole process; he was the Horizon guy. He talks in later paras about things like cheques, and having no idea what the process would be if there was a mis-match. He seems to have been the most narrow of "architects," designing a system that seemingly operated in some sort of vacuum.
Staggering.
Could you give a reference to that document you quoted ? I.e. where is it in the video, or on the website ? It's interesting if it really gives an obvious mechanism for the discrepancies (poor UI design leading to mistakes by a clerk).The implications are breathtaking. Have a read of the paragraphs leading up to para 27. I'll type out a bit here, which relates to daily and weekly declarations of cash and stock. Users were expected to *manually* use an ID number every day to denote a cash drawer and, if they typed the wrong number in (intending it to be common), they'd get a larger underage/overage.
Draft Witness Statement said:
As part of the Balancing process a similar cash Declaration is done using the Declaration IDs exactly like the ONCH declarations. In 2004, these were separate independent declarations, but in 2005 they were merged to use exactly the same process for both ONCH and Balancing declarations. As part of balancing all Declarations with separate IDs are taken and added up and then compared with the System Generated Cash figure which is produced by analysing all transactions within the cash account period. Any difference is highlighted to the subpostmaster as a <<Discrepancy>> which if the[y] accept, will adjust the system cash figure to match the amount of cash that has been declared by the subpostmaster. Therefore if he has <<doubled declared his cash>> and accepts the resulting discrepancy, the system cash figure will be adjusted to reflect this. I suspect that this may be what has happened. However it is still Mr Castleton's responsibility to spot this and not to accept the (false) discrepancy when he is balancing the system.
Wow. Please tell me the likes of GJ weren't in charge of what we'd now call UX (User Experience), and that genuine HCI [Human Computer Interaction] people were involved in what was then Europe's largest IT project. To rely upon SPMs and their (low wage) staff to accurately type in a till ID every day by hand and, if they get it wrong, to double-count all the cash in the building, is just insane.Any sensible UX person would have a configuration system that allowed the SPM to specify the tills, their IDs, the basis of counting (all drawers together, or declared separately), and so on, and then *prompts* to count "till X", not "please enter the till ID."
If this is indicative of the way in which Horizon was designed, I'm pretty stunned TBH.
The other thing that's clear is GJ wasn't architect of the whole process; he was the Horizon guy. He talks in later paras about things like cheques, and having no idea what the process would be if there was a mis-match. He seems to have been the most narrow of "architects," designing a system that seemingly operated in some sort of vacuum.
Staggering.
Gassing Station | TV, Film, Video Streaming & Radio | Top of Page | What's New | My Stuff