, 82 tweets, 14 min read Read on Twitter
Welcome back to court 26 of the High Court Rolls Building for the afternoon session of claimant cross-examination in the Horizon trial - part of Bates and others v @PostOffice group litigation

Fomer Subpostmaster Angela Burke is on the stand. Next up is the Fujitsu whistleblower Richard Roll who featured in the 2015 Panorama “Trouble at the Post Office"
Post Office barrister (POB) is Owain Draper.

He is addressing AB re her Witness statement (WS) which suggests she didn’t think that if she had gone to the lengths she had gone to, she wouldn’t have recovered her £150.
POB we are not criticising you for the lengths you went to to recover your money, but you didn’t know what was going on at Post OFfice and I am about to suggest to you it would have been resolved in your favour anyway.
POB takes AB to a BIM incident report which is a "business incident management incident report” written on the same date AB was apprehending her customer and taking her to his bank to get evidence of the Horizon failure.
POB taking her to an email attached to the report. It says “I have attached a list of transactions which POL will have to manually reconcile as appropriate” See that?
AB yeah
POB takes her to the spreadsheet attached to the email attached to the BIM. Do you see your branch ID there?
AB yep
POB £150 cash withdrawal authorised by FI and authorised receipt was produced but it failed due to known data centre issue.
POB assuming money was handed over the customer account was correct, but the branch will be down and this will need to be manually reconciled. The Fujitsu employee didn’t know that cash had been handed over.
AB but I’d already told the PO
POB but this is a Fujitsu employee
… looking at the data. And you see where it says on the assumption the money has been handed over - reconciliation will be required.
Then to show what PO is doing in response to this analysis from Fujitsu… [we go to another doc]
POB you see row 13 has been highlighted with your
… branch code….
AB you see the time of the transaction - it’s at 0826. When my branch wasn’t open
POB the time stamp doesn’t take account of daylight savings time
AB okay
POB it just shows GMT
POB This is what PO says about this transcation. “Action taken: branch contacted, shortage in branch so TC issued to adjust cash” so that is PO with info it had confirming it will issue a TC. This coincides with your WS where you say your husband was told to expect a TC
AB yes
POB and here we can see the TC was issued on 16 May [the incident happened on 9 May] “to correct comms failure online for £150 with ref - so credit to office"
POB so that corresponds to your TC which arrived on 17 May
AB yep
POB so what I am trying to establish is that having seen now what you’ve seen, with the benefit of that knowledge it looks as if PO would have been able to resolved the problem in your branch even if you hadn’t...
… taken the steps you did.
AB possibly
POB no further questions.

Richard Roll up next.
Patrick Green QC for the claimants has alerted the judge to a PEAK (Horizon error) which the Post Office have only just disclosed. He wants 10 minutes to discuss it with Richard Roll before he starts giving evidence. J grants it. We resume at 1430.
In fact PG asked for 5 minutes. The judge, in full Simon Cowell mode, said: “I’m not going to give you 5 minutes. I’m going to give you 10 minutes.” Oh the drama.
And we’re back. Mr Richard Roll is being sworn in. Mr Roll worked at Fujitsu on Horizon support between 2001 and 2004.
Mr de Garr Robinson for the Post Office on his feet to cross-examine RR.

PO QC wants to ask RR about his role in 3rd and 4th tier support for Horizon.
PO QC you know Mr Parker from Fujitsu who has made 3 witness statements to the court.
RR I know he’s made some
PO QC have you read them
RR I’ve read some of them
PO QC closely or briefly?
RR briefly
They go back to way the support for Horizon is structured.
POQC 1st line is run by PO?
RR I don’t recall - I believe there were splits between PO and Fujitsu on first line, but I really can’t remember.
POQC Mr Parker says vast majority of discrepancy referred to us as pseudo-software issues are not - they are user error not Horizon issues. Do you agree?
RR No. My recollection is not like that.
PO QC can you read this description of the second line of support [I can’t see this document - it’s a Fujitsu employee witness statement which is referred to in the claimants’ opening statement]
RR yes
PO QC is that right
RR not my recollection - we had nothing to do with 2nd line - we ran 3rd and 4th line...
PO QC reading out what 3rd line support did. It is a long read.
Do you accept it is true?
RR yes
PO QC reads out 4th line support - it is a description of the super top peeps who write the software etc. Do you agree this is right?
RR broadly speaking yes
POQC you describe yourself as 3rd/4th line support when in fact you were 3rd line
RR yes
POQC and 3rd line are elite, 4th line are super elite
RR there were some 3rd line I’d describe as super elite to
POQC were you?
RR no
PO QC how busy were you
RR peaks and troughs - sometimes we have 3 or 4 jobs on the go at the time sometimes quieter
PO QC it says here it would be rare for a software error to go through to 4th line that needed fixing. Correct?
RR yes
PO QC would it be right to say of the calls coming into third line - a significant number would be routed back out
RR calls would come in - we’d work on them, either fix them or pass them on to other people depending on the nature of the problem
PO QC so eg an infrastructure problem would go to infrastructure team, not 4th line
RR probably
PO QC goes to “Live PEAKS into SSC” tab (SSC = Fujitsu Horizon Support Centre - System?) - this is from Jan 2001 all the way through to 2004. There are 25K-odd calls going to 3rd line support during that period. Correct?
RR Sounds about right...
POQC of the calls coming into 3rd line support - 14% went elsewhere and only a subset would have gone to 4th line support and only a subset of that would be software errors. In a position to challenge that?
RR no
POQC Mr Parker says only a tiny fraction of calls would be sent to 4th line support and only a fraction of that would require changes in software.
RR when you take, as an average then yes, I suppose so...
RR when I was there the nature of the work changed over time - the things that stuck in my mind was when there were difficult periods involving software issues.
POQC software problems requiring a software fix were a tiny fraction of the work of 3rd line support.
POQC were you Support Assistant?
RR I was a Product Specialist
POQC quite junior
RR most people were Product Specialist two or three were senior and there were about 25 of us who were Product Specialists
POQC I am going to suggest there were 5 top people and 25 junior
RR junior doesn’t really do justice to the level of knowledge required
PO QC I agree - ordinary level, but there were some really top people who were very specialist
RR there were guys who were more focused
RR they knew UNIX etc
POQC but what I’m saying is that they did the harder work
RR not sure how to explain - there were some jobs where I would be more experienced, but others where they would be far superior to me
POQC it would be relatively rare for more junior people outside that 5 to do the programming changes
RR well some of us were C programmers and we might look at a problem and then pass it on to the top level.
POQC you were well-trained werent you
RR yes by Fujitsu
POQC and there was a 6 month mandatory training period before they’d let you loose on the live system
RR I dont’ remember
POQC you did a lot of mundane tasks
RR yes - it’s the unusual and difficult tasks which stick in the mind. It might be that with memory I have thought I was doing less mundane tasks that I actually was.
PO QC that’s very fair. Let’s have a look at some of the tasks you might have carried out.
[shows RR a Fujitsu document]
POQC what does this task involve?
RR describes basic recollection of the task
POQC and this [another doc]?
RR can’t remember
POQC what about this “check for marooned transactions” what are they?
RR we had an end of day marker and any transactions done after the marker so some transactions may not have been harvested properly
POQC [interrupts] can I suggest to you these are marooned ..
… transactions inside a machine that may need rescuing.
RR yep
POQC takes him through it show that it was basically a variation of what they both said.
POQC moves to year RR left - "OBC exercise... planned short notice close… pinpad return for triage services…"
POQC is it fair to say a lot of your work involved supporting engineers doing this kind of work.
RR don’t know what proportion of my work was this sort of thing I can’t remembe
POQC goes to WS. [I’ll see if I can upload the WS and link to it without missing too much]
RR Witness Statement 1: scribd.com/document/40180…

RR Witness Statement 2: scribd.com/document/40180…
Here’s a link to the embeds of both Richard Rolls’ witness statements on the postofficetrial.com website


I don’t think we’ve missed much...
PO QC is still asking about the nature of RR’s work.

RR some reports were looking for comms problems. Others would be looking for errors in data stores. The DS would be harvested every day and processed overnight. That would generate reports.
And we would have access to those reports.
PO QC so these continuously generated reports and if they generated a hit that needed to be identified and dealt with
RR yes and as time went on we were writing more programs ourselves which would help alert us to problems.
I think this is going to be a very long (and very possibly accurate) attempt to demonstrate to the judge that Mr Roll was not a very senior member of the Fujitsu team
PO QC you weren’t generally involved in source code examination
RR not generally, but I was involved in some
POQC so you wouldn’t be writing code
RR no we’d identify a problem and pass it on to 4th line
PO QC you weren’t involved in the support provided by the 4th line
RR no
PO QC so reading copious amounts of code was something 4th line did. not you
RR yes
POQC you weren’t reviewing much code
RR my recollection is I was, but it might be that the interesting stuff was magnified in my memory
PO QC given what you say in WS1 par 7 you have given us a slightly misleading impression of the nature of your job - you didnt’ regularly go through thousands of lines of computer coding did you
RR I wasn’t doing it regularly
PO QC what is your recollection
RR that I did...
… regularly go through 000s of lines of code, but now I’m not so sure
PO QC I’m very grateful
[I’m having trouble dealing with what’s going on in court because everyone is reading off a WS from a Mr Parker from Fujitsu which I am not allowed to have until he’s been sworn in.]

RR can’t remember much of what’s being put to him.
We’re on a short break till 1535
Okay we’re back. PO QC is going through a list of PEAKS which happened whilst RR was at Fujitsu. The PO have basically analysed everything that happened to put his WS’s into context. So far they haven’t contested any of the Horizon errors described in his WS but...
… they are attempting to show how relatively rare they are - whereas RR’s WS gives the impression of an ongoing fire-fighting operation.

Seems a fair approach.
Essentially it doesn’t really matter how rare or common or otherwise they are - it’s the effect they had on the branch accounts of Subpostmasters. If any SPMRs were blamed for Horizon errors, or worse prosecuted for shortfalls generated by Horizon errors...
… it doesn’t matter if there were only 2 or 2000 errors a year.
PO QC is asking about what percentage of his time was spent working on KELS and PEAKS etc with ref to a document which the PO has created documenting what percentage of his time was spent doing what
RR says it’s not his recollection but the figures are the figures.
POQC so you’re not in a position to contest them
RR notes some problems he would have worked on and then passed to someone more qualified would not be shown on the document.
[the doc just shows the problems he had “ownership” of]
PO QC goes back to par 11 of RR’s WS: “I would describe much of the work as fire-fighting”
PO QC that’s wrong isn’t it.
RR on the figures here you are correct…
it’s a tiny amount of the work you did, isn’t it
RR volume-wise no, but obviously you spend more time on a bigger prob
PO QC I suggest to you Mr Roll that even when you had a coding error that created a data problem in a branch the number that created a branch shortfall was a vanishingly small proportion
RR my recollection it was a big part of my job so it didn’t seem like a vanishingly small
proportion to me. And also if we were passed a problem and they couldn’t find a problem it didn’t mean there was a problem. So I can’t agree fully with what you said.
PO QC [asks about doc which says coding errors were extremely rare]
RR yes
PO QC and that even 3rd line calls...
… would usually be down to user error.
RR my recollection was that it was pretty balanced between data corruption and user error
PO QC so 50% was user error?
RR maybe - not the majority was user error but it is possible
PO QC lets move on… [keeps reading from a doc by a Mr Parker from Fujitsu which is being used to rebuf RR’s evidence]
PO QC asking if they couldn’t find an answer to a problem it had to be user error
RR says sometimes we had a deadline and if we couldn’t find the source of a problem we’d have to leave it
PO QC how often
RR very infrequently.
PO QC looking at the figures for the time you were there 1.5% of issues going to SSC were receipts/payments mismatch - so less than 1.5% of issues would be causing issues in branches if you accept the vast majority of in-branch problems were caused by mismatches
RR agrees
PO QC corrects himself to say “potentially” causing issues rather than actually causing issues.
PO QC you start by saying the issues with coding in the horizon system were “extensive” would it be fairer to say it was not remotely extensive at all - in fact it was tiny.
RR on what you’ve shown me, yes
PO QC and you weren’t working on these daily were you?
PO QC it was actually a very small proportion of what was being done.
RR yes on bugs
J by bugs you mean coding
RR yes
J I seem to be the only person who cares about the exact language being used
POQC now on par.11 "My recollection is that the software issues we were routinely encountering could, and did, cause financial discrepancies at branch level, including “shortfalls” …"
“… being incorrectly shown on the Horizon system. If we were unable to find the cause of the discrepancy then this was reported up the chain and it was assumed that the postmaster was to blame."
POQC so management wanted you to find out what happened
RR yes we were under pressure
POQC and are you saying you didn’t have as much time as you liked. An in that situation you’d report it up…
RR yes to Mr Peach
POQC and what would he say?
RR no idea what he’d say to the PO
POQC did you come across any error which caused a discrepancy in a branch account
RR i did suspect some and passed it on, but I don’t recall uncovering one
[I think we’re winding up now]

POQC quoting Mr Parker docu "on the very rare occasion software error could cause an impact on branch accounts, Fujitsu would tell the PO and work on a resolution." Is that something you can comment on?
RR no
We’re on to housekeeping.

J updates on judgment. He has a list of typos from claimants and assuming PO has similar and so is on track to hand down judgment at noon on Friday.

Judge says judgment is 180,000 words long.
Friday becomes the biggest day in this group litigation so far.
@threadreaderapp unroll pls
We start again at 10.30am tomorrow for the rest of RR’s evidence and that of Ian Henderson director of Second Sight.

Transcripts, witness statements and report will be published on postofficetrial.com tonight.
If you want to join the secret email brigade, please consider putting a few quid in the tip jar at postofficetrial.com or paypal.me/nickjwallis

I would be very grateful.
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!