, 104 tweets, 17 min read Read on Twitter
Morning. We are in court 26 of the High Court’s Rolls Building for Day 7 of the Horizon trial - part of the Bates v Post Office group litigation. Third day of Post Office witness cross-examination. Follow this thread for live tweets.
*cracks knuckles*

Judge (Mr Justice Fraser) is sitting we are into housekeeping before the witnesses are called.
Please remember these tweets paraphrase and are summaries of what is happening in court. Nothing is a direct quote unless it is in “direct quotes”…

discussing the possibility of a day’s sitting next week to do some mopping up...
Post Office’s David Johnson (DJ) in the witness box. Mr Johnson is a training and audit advisor. We have been given his witness statements by the Post Office solicitors. He is about to be cross-examined by the claimants’ QC Patrick Green (PG).
He has demonstrated Horizon to some “representatives” in court today - I think that suggests he has show some of the lawyers how Horizon actually works...
PG taking DJ through the Horizon screenshots he used to illustrate his witness statement.

Like so many Post Office witnesses he is a lifer. Started as a counter clerk in 1984.
PG asking which screenshots have been taken from a how-to Horizon guide for SPMs and which haven’t. DJ isn’t sure.

PG has noticed that some of the screenshots aren’t from the guide by spotting some text was in italics on some screenshots and some weren’t.
PG asks if DJ put these screenshots in his witness statement.
DJ didn’t
Judge asks who did
PG doesn’t know

We move on.
Sorry *DJ* didn’t know…

We’re now onto buying a stamp. Again.

This time the physical interface procedure on Horizon is being discussed and they are going through it.
Judge interrupts to say to PG

“Exotic though your hand movements are, they’re not going to appear on the transcript.”

PG is now asking questions with his hand behinds his back.
PG noting how close together the buttons on Horizon, yet how much space there is on the touch screen.
PG they could be further apart
DJ it appears so yes
PG and you are aware of problems that SPMs (subpostmasters) have with miss-keying?
DJ yes
PG now looking at a internal Post Office document which notes Buy and Sell icons on travel money screen are too close together
This is a problem as the PO document says the situation cannot be reversed if the wrong button is pressed and the customer leaves before it is noticed.
DJ disagrees - he thinks both on legacy Horizon and current Horizon can reverse these - contradicting PO report.
PG goes on to 0 and 00 being so close together. Potential for misskeying?
DJ not really
PG says - well we’ve seen lots of mistakes made with extra zeroes on the end - is there not a potential?
DJ accepts
PG walking through why busy people make mistakes
PG explains that improvements are all about helping people stop make mistakes and asks him to agree that this is a good idea
DJ agrees
PG notes that the 0 and 00 buttons have been next to each other since Horizon started.

[weirdly doesn’t ask why it was never changed]
[we move on to polling]

PG you must make a cash declaration by 6.55pm - what happens if it goes at 7.01pm?
DJ cash management team use that info to work out what cash to send or collect from a branch.
PG SPMs cannot export data from Horizon can they?
DJ no
PG do you know if any thought has been given to this?
DJ no
PG Post Office has a lot more tools available doesn’t it
DJ yes
PG why doesn’t PO restrict itself to the same tools it gives SPMs. Why does it need Credence?
DJ I don’t know
PG Why doesn’t the PO just work things out from till rolls
DJ I don’t know if it can print out till rolls like an SPM
PG okay.
PG lets go to a situation where Mr Bates, a lead claimant is worried there might be duplicate entries in his account. How does he work that out?
DJ i would advise to raise the issue with NBSC [helpline]
PG but if it’s their responsibility to find their discrepancies in branch...
… what do they do? We know there are till rolls. We have one - it’s about *that* long.
J intervenes to ask DJ the process
DJ explains the filtering process and then the eventual report which is printed out
PG if you’re looking for a duplicate you’d have to remember the ...
…. figures you’d seen.
DJ well yes… [blethers] I’m guessing… but… I’m not quite sure what you’re asking
PG Okay let’s say you get a discrepancy of say £1792. This could be a single duplicate entry or many…?
DJ it’s not the most user-friendly way of investigating.
PG that’s all I was really asking
J you say you can filter in branch
DJ yes
J that would only help if you knew what you were looking for?
DJ not necessarily - you can filter to a day if you know when the cash declaration misbalanced...
[we move on]

We are now looking at p6 of the Network Development Programme Simplification internal PO document.

PG “there are a number of branch ops/processes around accounting and reconciliation - they are unecessarily complext… lack of visibility” etc etc
PG do you agreee?
DJ yep
PG “ill defined and out of date processes…” is that fair
DJ my experience is that suspense accounts work
PG is that a reasonable view in that document or completely unreasonable
DJ no it’s reasonable
PG no further questions

POQC (Anthony de Garr Robinson) on his feet. how experienced a Horizon user are you?
DJ very
POQC it was suggested that the buttons in H are too close together - comment?
DJ the fact the buttons are so close together means mistakes can be made...
… but it’s easily rectified
POQC how?
DJ if you press the wrong button you get the wrong total so you’d notice it
POQC do you have a view as to keeping the layout of a system over time is a good thing or a bad thing?
DJ good
POQC why?
DJ you get used to where the buttons are
J has questions about his experience. was it all in Crown branches?
DJ yes
POQC is there are material difference in operating Horizon in a Crown or agency branch
DJ no

DJ is done. Andy Duggs (?) is up next.
Witness statement will arrive shortly to confirm.

Andy DUNKS is his name! He works for Fujitsu. he is an Information Technology Security Analyst. First Fujitsu witness in the box. And a new barrister at the front bench.
CB for claimant barrister
AD for Mr Dunks

CB how many extractions have you carried out in your 17 years
AD hundreds
CB how many people in your team
AD 3 of 4
CB similar extractions?
AD yep
CB takes AD to Mr Godeseth from Fujitsu WS. Mr Godeseth has listed the number of ARQ requests from PO
CB 700+ in 2013/4 why?
AD no idea
CB would it make sense that was during mediation scheme?
AD yep
CB 2016/7 - 400+ because of these proceedings?
AD possibly
CB 100+ in 2015/6 - is that normal?
AD don’t know
CB we don’t have figures before 2014 - Mrs Mather said there is a contractual limit of 720
AD correct
CB so that means you must be keeping a record
AD yes
CB so why did we not get given them
AD don’t know
J were you asked for them, personally
AD no

[we move on]
[judge isn’t going to like that…]

We move on to AD’s job.
CB really the only time records are audited is when the ARQ logs are extracted?
AD we call it audit transaction records but it’s not pre-audited
CB so it’s slightly misnamed
AD possibly
CB you only give evidence on...
… Horizon online in your WS
AD it’s probably misleading. I’ve done extractions for legacy Horizon too
CB so when you refer to controls on data are you talking about all the time or the data you extracted for this witness statement
AD just for this WS
CB do you always provide filtered data or unfiltered
AD filtered
CB always?
AD never given PO unfiltered data
CB what do you mean by data integrity
AD that is what I’m led believe means is untouched
CB so not tampered with?
CB no not actually touched
AD so integrity means directly comparable with the live data

CB lets look at what Dr Worden [PO IT expert witness] says. He says it is secure and tamper proof

AD where the data is stored I don’t know so I can’t answer that, but in….
… regard to the handling of the data it is good.
CB a “gold standard”?
AD yes
CB it’s important, right?
AD yes
[Mr Dunks seems extremely uncomfortable]

CB asks about the list of controls he put in par 6 for his WS. When you set out these 12 controls when you were drafting your statement, did you refer to any document?
AD no I wrote them from memory
CB in terms of those 12 controls...
… have they always been available?
AD yes
CB so anywhere we look these 12 controls would always be in place at any point in time
AD yes
[we go to par 8 of AD’s statement]
CB [reads] "there is no reason to believe the info in this WS is inaccurate…"
“….because of the improper use of the system.” What is the system?
AD is not sure
CB you don’t know what you mean by system?
AD I think I mean the data extraction system
CB okay - the process for removing data [he reads on]
CB “at all material times” what does that mean
AD there’s never been a problem when extracting data
CB in all your 17 years
AD correct
CB so you are not aware of any problems with the data extraction process ever?!
AD no. [corrects himself] well there may have been problems...
… with reboots etc but nothing that would interfere with the integrity of the data.
CB how do you know if the system wasn’t operating properly it wouldn’t affect the data?
AD because there are checks and balances. we’d be told by the system if there was a problem in the data...
… and we’d re-run it to make sure we got a clean copy.
CB is your statement on integrity a party line?
AD no it’s not its what I am saying

CB let’s go back to Mr Godeseth’s WS [Mr G’s WS agrees with AD]
We go to audit extraction client user manual for Fujitsu. AD admits he hasn’t read it for several years.
It explains to the reader how to get ARQ data from Horizon and give it to the Post Office.
AD can’t remember reading it
CB continues to read from it “the integrity of audit...
… data must be guaranteed at all times”
CB that wording mirrors the wording in your WS [goes on to give more examples in which the wording of his WS matches this document, despite AD a) saying he wrote his WS from memory and b) not having read the document for several years]
[he must have a very good memory]

[CB keeps reading from audit extraction client user manual.]
CB “a gap in a message sequence may suggest there has been a duplicate in the audit data” possible?
AD yes
CB but it says “may” so it might not be.
AD yee-es…?
J intervenes - he didn’t write this
CB making the point that the duplicates COULD be down to the extract process or could be in the underlying data
AD I don’t know
CB have you had an error message to that effect
AD can’t remember
CB it says here “seek resolution from audit support"
CB what do audit support do?
AD I can’t remember I guess they’re try to sort it out.
We’re on a break. I have just been told the claimants’ latest barrister rejoices in the name of Ognjen Miletic. The solicitors just told me they call him “oggy”. He is apparently something of a tech expert.
I’ll stick with CB for continuity.
We’re back.

Gareth Jenkins WS from 2010 for Seema Misra now being discussed.

CB I assume you are fmailiar with the Misra case?
AD yes
CB you provided a WS
AD yes
CB Ms Misra was a West Byfleet SPMR charged with fraud and theft
AD yes
CB have you seen Gareth Jenkins WS before
AD no
CB reads about legacy Horizon data extraction from GJ’s WS and notes that there has only been one instance of legacy Horizon duplicating data.
CB reads from GJ WS which notes a new retrieval tool on Horizon online was creating duplicated data from ARQ logs - do you remember this?
AD vaguely
CB so there was a semi-automated process for removing duplicated records generated by this tool. Remember that
AD not really
CB so that wasn’t something you were involved in that or anyone in your department
AD that’s right - that was not done in our team
CB goes back to GJ WS re Misra trial - and notes there were some duplicate transactions which made their way into info re the Misra trial...
…. but there is reassurance from Fujitsu that the data is accurate.
CB why does the language in GJ’s WS from 2010 exactly match the language in your WS?

AD we do have a standard witness statement we produce for ARQ proceedings

CB so your witness statement was just cut and...
… pasted from your standard WS document without much thought.

CB notes another paragraph in another witness statement in 2010 from the Misra trial. Again it is almost identical to a paragraph in yours.
AD yes
CB takes him to a section in AD's colleague’s WS from 2010 and notes the 12 controls in her WS. They are almost identical to the ones he wrote from memory for his WS. But with some slight differences. She doesn’t reference where her controls come from. And neither do you.
CB how did you satisfy yourself that the controls are in place
AD I was trying to imply the data I had extracted hadn’t been touched
CB no - I asked you how you satisfied yourself the controls were in effect when you extract the data
AD I had no reason to doubt it
CB even tho you didn’t confirm it by looking at anything
AD no
CB goes back to the Misra trial data extraction

Ms Tomasin not identifying duplicates is a problem for the Post Office and a SPM trying to get to the source of the problem isn’t it?
AD we just extract the data
CB and identifying gaps or problems is not your concern?
AD no
CB and you can see how gaps in ARQ can be a problem.
AD possibly…
CB are you aware of any other problems?
AD no
[we move on to a PEAK]
It is a PEAK about an ARQ data retrieval problem.

CB “spreadsheets provided the prosecution team missed out whether a transaction was a reversal” “prosecution team well aware” “fix must be got out or prosecution evidence is not completed” This is Aug 2011.
CB you say you weren’t aware of this or the changes made
AD I don’t recall it no
CB “fix deployed” “all is working well” so up until 2011 the data you retrieved did not have an indicator that information being extracted was incomplete. Is that a problem?
AD I don’t know...
… whether it would have an impact or not.
CB did you hear the Helen Rose report being mentioned on Tuesday?
AD no
CB it wasn’t until the summer of 2013 that the reversal indicator didn’t show whether it was system or user generated. Is that outside your knowledge
AD yes
We’ve just been through several issues that Fujitsu has extracting or relying on that data - none of which is in your WS. Why?
AD I was trying to tell you about the integrity of the data extractions process, not the data itself.
CB so you can’t say the data itself
handed over to the post office is reliable?
AD I can't, no.

POQC reads par 8 of his WS "There is no reason to believe that the information in this statement is inaccurate because of the improper use of the system. To the best of my knowledge and belief at all material times…"
…. the system was operating properly, or if not, any respect in which it was not operating properly, or was out of operation was not such as to effect the information held within it.”

and asks if he would have signed his WS if he didn’t believe it to be true.
AD confirms if he thought it were not true, he would not have signed it.
Mr Torstein Godeseth from Fujitsu is sworn in.
He has produced 2 WSs - second adding clarifications and corrections to the first. POQC asking questions which is odd. Normally it is a perfunctory “is that your witness statement”.
Big disco about duplicates.
Mr Godeseth seems to really know his onions. He thinks that the issue of duplicates is misinterpreted.

TG - A record can be written to an audit trail twice, and as long as that’s spotted and dealt with it’s not a problem.
POQC if duplicates were being produced in audit extractions - would Fujitsu want to know about that.
TG - does know his stuff - he wants to go to absolute brass tacks on how it works. Describes what an audit trail is. HOL has a journal sequence number which stops...
… missed data and duplicates. TG describes it as a Big Protection. TG describes the Riposte system protections in place to stop dodgy data being generated or produced.
Talking about Wigan and Bootle data centres and the copying of info.
J re-asks POQC question. If duplicates are being produced. Is that something Fujitsu would be interested in knowing about?
TG yes because we’d have to explain it
POQC are you aware of it happening in HOL?
TG no
PG on his feet. TG is a chief architect of Horizon. First employed by PO?
TG in 1987 for a number of years.
PG goes to par 5 of TGs WS…
TG’s full name is Torstein Olav Godeseth. He sounds like he’s from Somerset.
He is an Oxford graduate. He joined PO in 1987 at the IT department to run a Thames Valley pilot IT scheme affecting 150 offices (Horizon came in in 1999/2000)
PG and Horizon when it started was being procured by DSS and PO. So it was unusual.
TG that’s right my Lord
PG you were tech advisor to the PO when Legacy Horizon went to Horizon online
TG yes
PG then you went to Fujitsu in 2010. Why?
TG my contract came to an end at PO and I picked up a contract at Fujitsu. In November last year I accepted a full time role
TG I became an contractor to the PO in 2005 when the Impact programme was being launched.
[explains impact problem]
PG you have effectively worked in Post Office and Fujitsu and the whole Horizon project
TG not involved after procurement for a bit
PG clarifying your WS
PG broadly covers remote access for legacy Horizon and Horizon online.
TG yep
PG you speak in your WS about consulting with colleagues in areas to ensure you were correct. which areas
TG nothing specific
PG so you were sufficiently uncertain.
TG no just double-checking stuff
PG Gareth Jenkins’ name crops up a lot in your WS. The source of a lot of your info in your WS is GJ isn’t it?
TG I had a pretty good feel for it. I had to know how Riposte works, but it’s right to say I needed GJ’s input
PG listing lots of “I understand from Gareth”s
in PG’s WS. This is info you have obtained from GJ, isn’t it?
TG it’s difficult to judge how much I knew beforehand but talking to GJ freshened up my knowledge. GJ is the expert on Riposte.
PG he’s the person to talk to
TG yep
[we go to Balancing Transactions]
TG writes: "A small group of Fujitsu users from the Software Support Centre (SSC) (30 users) have ...
… the ability to inject additional transactions into a branch's accounts in Horizon Online, using a designed piece of functionality called a Balancing Transaction."
PG is 30 users the whole of SSC?
TG about that
PG let’s look at the words “a small group” means in this case “the whole of SSC” yes?
TG yes
PG takes him to "the node ID (see paragraph 38.2 below) associated with the injected message would have been that of the correspondence server at which the message had been injected and not a normal counter node ID and that would have been clearly visible in any audit data."
PG notes this is the same as par 58.10 which TG says he got from Gareth Jenkins
TG now clarifying that injections could be made at the counter.
PG was section 36 [which I’ve just quoted in the last tweet] from Gareth too?
TG yes
PG now going on to say that TG in his second WS clarifies that counter injections were possible. Who told you this?
TG steve parker when he was preparing his WS
PG were you shocked by this?
TG not shocked. I was just finding out things I didn’t know before
PG in quite...
… a controversial area.
TG it was done because it was an operational necessity in quite a controlled way. i had expected messages to be introduced using a different counter position. The standard practice was to label what we are doing so we could see it in an audit trail.
TG SPMs would never see it and I am fully aware that was the case. it was a better audit when Mr Roll said in his evidence it was just recorded in Pinnacle. The intention was always to say this had been done in exceptional circumstances.
TG we could have done this by inventing counters so if a branch had two counters we’d add additional counter numbers. Like 11 or 12.
PG the short point is you learned when Mr Parker was prepping his 2nd WS it was in fact possible to inject transactions which an SPM...
… would not know about at the counter rather than at the correspondence server.
TG I don’t know how they would know that these transactions weren’t being injected. Someone had to be in the branch logged on when these injections were made.
IF someone wasn’t logged on then RIposte
… would log it as an empty counter ID and we needed Riposte not to wrap the envelope in the payload we were looking to inject. and that blank user ID would have been picked up.
PG let’s separate what you can do not what you’ve inferred from PEAKs. You realised for the first time
it was possible to do this when Mr parker was preparing his 2nd statement. YOu knew that was a contentious area in this litigation. Yes?
TG yes
PG and if we go back to your WS - “in legacy Horizon any transactions inserted…” would be greater than 32 so it would be clear...
… it had been injected in this way.
PG it’s not clear that you got this info from Mr Jenkins. So you had a conversation with Mr Jenkins on a contentious point in this litigation. He told you something and you believed it. Then you found out it was wrong.
And did you have a conversation with him about why it was wrong?
TG no
PG did you not want to know how he had that information wrong?
TG I didn’t have a further conversation with him
PG so your knowledge it limited to what you can infer from the PEAKS
TG yes that’s fair
PG so apart from a 2 outsourced period to Zanza and a short break, you’ve been at the Post Office/Fujitsu since 1987.
TG yes but I wasn’t involved in the rollout of Horizon - I went to parcelforce etc
PG are you regarded as a knowledgeable person within PO as to whether remote
access is possible
TG knowledgeable
PG very knowledgeable?
TG very knowledgeable yes
PG and is your evidence to court that you were completely unaware remote access in this way was possible
TG if asked I would say it was inevitable in a system this size
PG so were you surprised when Gareth Jenkins told you it wasn’t possible
TG I’m not sure what he told me was that specific
PG but this was the point of your evidence in 58.10
PG that remote access was possible in a way that the SPM cannot identify
TG that’s not the point I am making
PG [reads him the point] you were specifically ruling out transactions in a way a Subpostmaster could note see
TG asks him to repeat the question
there is a long pause

TG I think there are too many editors in this

PG you say "In Legacy Horizon, any transactions injected by SSC would have used the computer server address as the counter position which would be a number greater than 32,…"
“… so it would be clear that a transaction had been injected In this way.” this makes it clear there is no way an SPM cannot see a transaction

TG doesn’t accept the point.

PG my lord shall we leave it there?

we do. for lunch.
@threadreaderapp unroll pls
Missing some Tweet in this thread?
You can try to force a refresh.

Like this thread? Get email updates or save it to PDF!

Subscribe to Nick Wallis
Profile picture

Get real-time email alerts when new unrolls are available from this author!

This content may be removed anytime!

Twitter may remove this content at anytime, convert it as a PDF, save and print for later use!

Try unrolling a thread yourself!

how to unroll video

1) Follow Thread Reader App on Twitter so you can easily mention us!

2) Go to a Twitter thread (series of Tweets by the same owner) and mention us with a keyword "unroll" @threadreaderapp unroll

You can practice here first or read more on our help page!

Follow Us on Twitter!

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just three indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3.00/month or $30.00/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!