PG and doing so is open to error?
PG and were you aware authorisation from the PO to do so was needed
J “in the happy days when £484 could by $1000”
PG… and the solution is to make a change..
PG "it was fixed before they did the roll and it is best that the branch is not advised"
PG so Post Office will have been aware of what was going on
PG the counter problem has been corrected by inserting a message into the message store...
PG by using an OCP
[we are getting techy here…]
PG single SC message was written in error with no corresponding settlement line - we will insert a new message. If the change is not made before rollover the branch will have an unexpected gain and a ...
PG so the OCP is the front end fix -
going into the counter message store at the branch
PG we see location: counter
PG so what we are seeing here is $2000 being inserted in the PO system. Yes?
PG when we go back to the underlying PEAK on 14 Dec 2007...
PG is it more likely that the loss has been generated by the insertion of the $2000 dollars?
[THERE YOU GO! THAT’S THE SMOKING GUN! $1000 dollar loss in branch blamed on SPM with a Fujitsu witness in court agreeing that it was more likely to...
[I wonder how often this happened…]
In court we have moved on to another PEAK. Once TG agreed that a branch had been blamed for a loss by a mistake at Fujitsu...
PG reading email re Malformed Currency Transactions between Fujitsu and Post Office “POL are happy for you to make necessary adjustments” - this is Post Office authorising more tinkering with branch accounts. This time WITH SPM permission.
(read it here postofficetrial.com/2019/03/horizo…)
PG where did you get this info from?
TG the guys
PG which guys
TG steve parker
PG and it came to him…?
PG so here they double the receipts from £4000 to £8000… and there it says an OCP approved by POL will be needed to do it using the Transaction tool
PG why couldn’t the £4000 be corrected with a Transaction Correction to the SPM
PG this was the planned way to do it
TG well we planned never to do it.
PG here we have a PEAK showing two transactions with the same JSN number
TG not sure this PEAK is accurate -it’s just the diagnosis of a developer
J let’s just go on what we have on the screen. You are not agreeing with the diagnosis, but let’s go with the...
TG it would have to be a bug in ORACLE because that is an impossibility. And I certainly don’t remember any such bug.
J I think the whole case it fairly complicated, but I think I can follow it. Carry on
TG [gives a very long explanation about a Horizon bug causing problems around the launch of Horizon online]
J I understand.
TG someone may have been trying to look for a JSN entry - I’m simply asserting an ORACLE bug which would have been huge...
[we move on to a PEAK in April 2010 - a month after the acknowledged balancing transaction you refer to in your statement. Problem here is FAD code [branch] unable to rollover]
PG then Gareth Jenkins suggests a different solution -
TG using the Balancing Transaction tool to delete an entry in the branch database which is NOT...
PG there are, I think two things happening here - there’s a stock unit entry being updated AND a deletion to the opening figure of cash.
TG I must admit I hadn’t picked up on that nuance
J I’m not making a finding here, but it’s not exactly a nuance if there are two different trading period
TG apologises and says that wasn’t what he meant
PG reading out loud from this doc - one of the benefits listed is: "fix which may not be visible to end user”
PG so it’s anticipated this tool would continue to be used going forward
PG quotes "when we go off-piste” - what does going off piste mean in this context
TG creating a fix we don’t have a script for
PG on 9 Dec 2011, Ernst and Young asked Fujitsu about privileged access to branch transactions...
PG here is letter from WBD (PO solicitors) to Freeths (claimants’ solicitors) about usernames in Fujitsu with a list of usernames and who they relate to. PG lists them.
There is also a letter listing what privileges they have...
TG those people could log on and do that yes
PG and prior to 2015 the only audit available was to log on and log off
I have interviewed both these people for the BBC in the past.
PG and that he doesn’t know anything about it
PG and that it’s only in your WS because you were asked to put it in there.
TG [pause] yes.
PG it which affected Callander Square in 2005, but was traced back to Feb 2003 and there were iterations of similar bugs in 2000. Correct?
TG [small pause] yes
PG reads out that Post Office helpline and told SPM there were no software bugs...
PG but there was a bug, wasn’t there?
TG I think we’ve established there was a bug, yes.
PG this defect which was discovered in 2005 and fixed in 2006. It wasn’t, was it?
TG no it wasn’t - the underlying bug had been there since the launch of Horizon.
PG that’s also incorrect, isn’t it?
TG there is a spreadsheet which lists the impact and how many might be impacted
PG it’s wrong, isn’t it?
J you might want to look at one of the PEAKS which PG took you to [he takes him to it] “a few of these errors seem to occur every week at different sites” does that help you answer Mr Green?
TG I would read that as a fairly generic problem...
PG takes him to his WS in which he says he spoke to a Matthew Lentern who told him there were 30 branches affected...
TG I didn’t speak to Matthew Lentern about it. so no.
PG did you have any communication with Matthew Lentern at all?
TG he gave me the spreadsheet
PG you’ve always known that the source of the info that you talk about re Callendar SQ in your WS was this Anne Chambers spreadsheet.
TG accepts, but didn’t originally know AC put it together
PG let’s see what...
PG says the spreadsheet shows the MAJORITY of branches affected and that it doesn’t know about how it affected other branches going forward. This does NOT reflect your WS, does it
TG fair point
PG you also said it was made at the time of the error, but the
TG yes I got that completely wrong
J was it the number of branches shown that led you to correct the number of branches affected from 20 to 19?
TG someone else noticed it.
TG confused about whether it was Horizon online (HOL)
PG let’s look at your WS
TG oh yes it was HOL
TG I have not checked the detail, but I’m comfortable with it
PG is this something you know about yourself or is this something you got from Gareth Jenkins?
TG it’s from Gareth, plus further...
PG and the doc you refer to in some detail is f1000 written by Gareth Jenkins. and doc f1001 is the receipts and payments mismatch issue notes which you are not involved in and don’t refer to.
PG then mentions another doc, and the BBC broadcast on 7 Feb 2011 re Horizon [which I fronted] and says Gareth didn’t tell you about those, either did he?
Monday 7 Feb 2011o
PG reads from a document where the Post Office decide not to tell branches about it. were you surprised about this?
TG it was the Post Office’s decision
PG do you agree with it?
PG [provides documentary evidence that the bug had been going since May 2010.] so your WS is materially...
TG on the basis of what we’ve just been through, yes it is.
[the thing that still gets me is that the Post Office could have told the Misra trial about the Receipts/Payments mismatch bug whilst it was happening, but chose not to. Now - they...
PG asking TG to explain how he came to the number that 60 branches were affected by the bug
TG can’t explain where it came from.
TG says he remembers Gareth
PG goes to the spreadsheet which shows how many branches were affected.
[This is a real Horizon bug causing discrepancies in branch accounts and SPMs not being told about them.]
PG now asking about the PO’s strongly held assertion...
PG do you know anything about that?
PG does that suggest there HAS been more than one Balancing Transaction injected into branch accounts
Now the claimants QC and his team have found it. How many more were there, I wonder. Back at 2pm!
Judge says he’s completely comfortable with this, but we’ll review at the end of Mr Godeseth’s evidence.
TG we found out about it in 2013
PG because Subpostmasters who brought it to your attention
PG and the Post Office knew about it in 2012
TG I can’t comment on that.
[we go back to the Dalmellington bug]
PG so there’s a 2010 fix to this bug and a 2011 fix
PG listing the incidents of the bug repeating which continue up to 2016. The point someone making at Fujitsu is making there is that the PO didn’t...
TG I can’t comment on that
PG what does it look like
TG it looks like it, yeah
PG so no root cause analysis could be done on it by Fujitsu for 4 years from 2011 to 2015
TG it’s possible Fujitsu did not find about this until 2015 and it’s possible that SPMs were dealing with it themselves.
PG it does not show Horizon robustly working as it should
TG the bug had the effect of making the user...
[which doesn’t answer the question]
PG finally - the assessment appears to be that Horizon has improved over the years
TG any system improves over the years
PG what about Horizon?
TG yes it has improved
PG this bug could cause a branch to be held liable for this bug. yes?
TG you’d get a receipts/payments mismatch which would be picked up by our system.
PG is that a fair reflection of the sort of thing...
TG I am not familiar with that bug.
[that didn’t sound like an answer to the question but judge and QC let it lie]
PG no further questions
POQC on his feet for re-examination.
TG that would be ridiculous because you roll out to a small number of branches to go through the...
POQC and the balancing problem that doubled the £4000 to £8000 that we heard about is that the sort of thing you’d expect
TG I’d expect to have to deal with a number of problems.
TG the raw data will not disappear.
POQC what is the effect of the delete instruction
TG to delete the data from the branch, but not the underlying data
POQC and what happened...
TG we needed to remove a line of code which had an extra zero in it. The zero caused the data to be left in the transaction data after rollover and that confused the software.
POQC and so by deleting the record what happens?
TG it allows the rollover to happen.
TG fundamentally you’re logging on to the branch database
POQC who is UNIX user?
TG one of the guys...
POQC let’s see this name here - Ed Ashford - who’s he?
TG one of the guys in Ireland
POQC who are the guys in Ireland?
TG they are UNIX users operating at a very low level
POQC what does that mean - they don’t have very high user rights?
TG I think the data that was deleted was NOT transaction data.
J what do you mean then in this case by transaction data?
TG well in this trial we’ve been calling it operational data, but that’s such a wide definition.
Any links to any precedents would be useful too.
I’ll put up the transcript at postofficetrial.com when I get it.
This judge seems to know quite a bit about binary variables.
Mr Godeseth leaves the witness box.
I am going to give you (POQC) some directions on your application.
J I received the application at five minutes to 2 and asked for a paper copy to be brought to court
J I think there might be some exhibits. Are there exhibits?
[a file is handed up]
1) responsive evidence from claimants and if they want to oppose it
3) date of hearing
Claimants’ QC (PG) on his feet.
PG if you lordship would turn to p6 of Mr Parsons Witness Statement. The basis of the application is supposed to be set out in par 24 and 25 - the findings in relation to matters to Common Issues trial - show you have...
And the high level of invective directed at the Post Office witnesses suggest a hostility to the post office.
PG I have not seen a recusal application as lacking in particularity as this. They need to be set out properly….
this is likely - if not calculated - to derail proceedings.
PG I am inviting you to direct that these vague allegations be particularised. No solicitor would ever sign...
J well I’ll tell you what I’m going to do.
Post Office QC - my lord, neither have I
J I’ll come back at 4.40pm and tell you how we are going to proceed.
This application to recuse the judge has been made by the Post Office.
The claim is that his judgment and the language in it is so strong it is de facto proof he cannot NOT be biased going forwards in this litigation.
POQC saying that claimants demanding particulars is surreal - it is a matter of argument. We suggest out skeleton is issued on Tuesday, the claimants is issued on the end of Thursday which will allow you hear the application on 1 April [apt]
POQC says a particularised list of evidence would slow matters down
J if the PO’s skeleton includes a schedule about the "critical invective" and the “harsh
J as I understand the WS of Mr Parsons which I have now read carefully twice...
POQC I don’t know
PGQC on his feet. for the claimants.
PG the starting point for any application of this type is that it must be set out with ...
We have to be given evidence of this claim BEFORE we decide whether or not to oppose it.
We have to see the reasons
A finding which is advers to a party is not the proper foundation of an application.
A criticism of a witness is not the proper foundation of an application of this kind to the court.
So it needs to be set out EXACTLY what creates a situation whereby the claim of bias can be considered.
J what I am minded to do is order the PO provide a WS by noon on Tuesday which...
J I think we’re all agreed we need...
POQC that’s not fair on Mr Parsons - a human being criticising a High Court judge.
POQC it would be much better if the arguments were spelled out...
J disagrees and sets out schedule:
Noon on Tue 26 for Witness Statement.
Claimants have to decide whether or not to oppose on this basis. They have to be given the oppo to put evidence, but I can’t imagine that they would if the basis...
So PG when would you like to notify the defendants if you choose to oppose the application
PG 24 hrs
J okay noon on Wednesday and response skeleton by Friday.
Skels by 10am on Tuesday 2nd
Application hearing will last 1 day and I am proposing...
POQC my lord I don’t know if Lord Grabiner is available the day
J I think then if you want to shift that date, given that the date is a day the trial should be sitting you’re going to have to make a formal application.
Judgment might be reserved. Judgment might also be appealed.
J says we now won’t be able to hear any expert evidence this side of Easter.
Judge rises. Wow.
Here are some excerpts from Mr Parsons’ supporting Witness Statement:
wide range of matters which properly fall to be decided at the Horizon Issues Trial
or at future breach trials."
training, the quality of helplines, Post Office’s alleged knowledge of problems with
Horizon, the cause of shortfalls, how easy or difficult it was for Subpostmasters to
discover the cause of shortfalls,..
circumstances of individual Claimants’ suspensions and terminations, and
whether Post Office sent unjustified demands for payment and/or threats of legal
action to Subpostmasters."
"These findings give the clear impression that the Judge has already formed a firm
view on these matters. It is to be expected that this will prevent him from taking an...
with the benefit of full evidence and disclosure."
Office, none of which is relevant to the determination of the Common Issues.
That, too, creates a clear impression that the Judge has not behaved impartially….
Office’s witnesses on matters irrelevant to the Common Issues."
application for the Judge to recuse himself from these proceedings. As an adjunct
to that, Post Office applies for an adjournment of the ongoing Horizon Trial."
One lawyer told me...
If it fails, I can only see the rest of the trial happening in the weeks originally scheduled for the third trial in Nov 2019. But that’s a guess. But these QCs...
I see @ComputerWeekly have filed already.
I’ll get something up alongside today’s transcript tonight.