Thread by @nickwallis

We are back in the High Court Rolls Building court 26 for the afternoon session of Day 4 of the Horizon trial - second trial in Bates and others v Post Office.

Richard Roll is still on the stand we are talking about remote access. Mr Roll is telling the court how they would access a computer’s info to fix it.

POQC you would never and never manually changed a line of transaction data
RR we did when we changed a line of code, but we’d never change £10 to £100
POQC there is no way anyone at Fujitsu would change any transaction data
RR that’s not my recollection - we could do it and did

change it without the SPM knowing it was us.
POQC no one at Fujitsu would do this - it would be more than they’re job's worth to muck about transaction data
RR we were not mucking about. without correcting a corruption and re-inserting it, it would continue to cause problems.

[RR has been asked to read something from Mr Godeseth’s WS]

POQC is that what you’re talking about?
RR no i don’t recognise this - the data could have been transaction data that we were changing
POQC you say you’d contact the branch...

.. ask them to not use it, then we’d log on. You wouldn’t actually log on to it, would you?
RR yes the terminal had to be switched on - we would go in with our remote log on - and this would be noted in the logs.
However for some transactions we’d have to go in with...

… their ID in order to satisfy Riposte that this was the SPMR ID.
POQC when would you do this exactly?
RR I can’t remember specific reasons we would do this. If we were doing it 2 people would have to be there to be an independent witness and it would be logged at our end.

POQC so this was a formal process Fujitsu took very seriously.
RR yes
POQC and the PO had to give their consent?
RR we were supposed to get the SPM’s consent - I’m not sure we’d go to the PO for consent.

POQC asks him if he is familiar with a protocol which would alert the PO
RR no
POQC and you’d have to get permission from the SPM
RR the idea was we should and we did
POQC so when you did this proper protocols were followed.

POQC the process of manually altering branch accounts and transaction data alterations had strict controls over them, didn’t they?
RR yes

[they move on to corrections to SPM terminals without SPM knowledge]

POQC any changes you make without SPM awareness would not affect branch accounts?
RR correct
POQC any examples?
RR I can’t remember

[they move onto cyclic redundancy checks - as neither a techy or legal person AND someone who doesn’t have sight of half the witness statements which are being referred to in court I am struggling. And I’ve got a cold. Poor me.]

RR seems like a new man this afternoon. He was being pummelled this morning, but seems on much firmer ground this afternoon on the remote access front. Sounds like he did it a lot.

POQC is it right that Fujitsu would be extremely reluctant to make changes to the basic features of transaction data - it wasn’t their job
RR it would have been our job because it was trying to fix a corruption in the data
POQC it was only done when you HAD to do it?
RR yes

[on to 2nd WS par 20] "Dr. Worden refers at paragraph 1114 to the possibility of SSC injecting transactions in old
Horizon, which is something I confirm that the SSC could do and that I did do this during my
time at Fujitsu. …"

“… However, Dr. Worden relies on the evidence of Mr Godeseth who says that any
transaction injected by Fujitsu would be “clearly distinguished” by a counter position “greater
than 32” because it would have been inserted at the correspondence server. "

“… However I think
there is a very important factual point here, which is that whilst we could insert transactions
using the correspondence server (and these would have a counter position greater than 32),
we could also insert transactions remotely by using the correspondence...

… server to piggy back
through the gateway, to the individual counters in a branch. The nature of many problems
meant that we had to implement a fix in this way rather than going to the correspondence
server, and we frequently did use this method in practice."

POQC can you explain how?
RR no - but it happened quite a bit

WS: "Sometimes we had to ask for a specific person to log in to the counter before injecting
transactions so that the software would not detect any discrepancies.”
POQC explain why that was.
RR if Riposte was looking at certain transactions going thru..

“… and it saw a different user ID it would reject the transaction.”
POQC we’re not talking about transactions in branch are we?
RR yes
POQC you said you didn’t
RR i misunderstood - we did. I wouldn’t insert data to correct the system when it hadn’t been written...

… but we would correct it. Certain members of SSC did that.
POQC but you didn’t?
RR not me no.

POQC your earlier evidence - you were asked about that and you said you didn’t do it.
RR you have two types of data here - the first one yes, the second one no. The first one we corrected corrupted data with the SPM logged on to rebuild message store on the counter...

POQC I think what you’re describing is a) get hold of the data that hadn’t replicated and b) find the line of data and correct
RR just to stop you there - we had to do that the other way round
POQC was that line of data a new transaction
RR no it was an existing transaction..

RR but there was the facility to create a new transaction with a new line of code. [goes into long explanation of how he “believe”s this happened]

POQC and RR agreeing on how he inserted transaction facility in a way that had an impact on branch accounts - as a corrective.
POQC you says it happened every couple of months? Could it have been less?
RR yes.

POQC on the occasions when you had to piggy back an SPMs login - am I right in saying the default procedure was to do it through the front door.
RR yes - further away you got the riskier it was
POQC and Fujitsu didn’t like taking risks. that’s not what you were there to do?

RR correct.

[POQC taking RR to a PEAK from 2009 long after he left]
POQC asks RR to explain a harvester exception
RR when the overnight UNIX servers the harvesters would run sequentially from 2am to grab the massive amounts of data from the Post Office terminals.

POQC now quoting from Gareth Simpson report about this harvester exception.
RR GS new more about UNIX than anyone there.
J intervenes to ask why RR is being asked about a PEAK 5 years after he left Fujitsu.

We’re on a break. Mr Roll is a changed man this afternoon. Some interesting detail in much of his evidence, compared to this morning when he really didn’t seem sure about large chunks of his WS.

We are back. We are undertaking what the POQC is calling an expedited xe as we still have Ian Henderson (@forensicgod) to do his one hour on the stand.

This will overrun I think.

RR being asked to make comments on the 2009 example and RR pushes back saying the 2009 version is very likely very different from the 2004 version.

RR being asked if as well as noting what they were doing wrt to remote access in their own “Pinnacle” system, they could also leave some note in the Horizon system.
RR isn’t sure.

RR my recollection if we were rebuilding the message store then we would go in but we wouldn’t leave any notes.
POQC how clear is your recollection
RR it was 17 years ago so… I might be wrong
POQC that’s very fair of you

[onto RR’s WS2 par 23] "As part of my role in the SSC, I was involved in rebuilding branch transaction data where a
particular counter became corrupted. "

“… Whilst in general terms I agree with Dr. Worden’s
summary of this at paragraph 1131, his description is very much a simplification of the process,
which was not straightforward."

“… When data on a counter became corrupt, the effect was that
data transmitted after that corruption could become stuck on the server for that counter and
not be replicated to other counters, the gateway server or correspondence server…."

“… In order to
correct the problem, we would have to copy the data after the corruption to our own computers,
fix the corruption and then replace the data with our corrected copy and build this back on to
the counter server."

“… There was clearly room for error in this process, where data could be lost,
or mistakes made when replicating data. As part of the rebuilding process, we were able to
delete data from the counter server, and did this as part of the process I have described above."

“… It was therefore possible that this process could result in data not being accurately copied
across and/or data being permanently deleted. Data was occasionally lost because we were
unable to copy it form the counter and thus were unable to rebuild the database correctly."

POQC pointing out all the checks and balances in place to minimise human error in re-writing that code, including 2 person sign-off. Why didn’t you mention this?
RR because it didn’t seem important, but I see how it might be

POQC you had someone literally watching over you while you did the work..
RR not always
POQC when then?
RR when it came to re-inserting it

POQC suggests mitigation steps taken to minimise errors into being “infinitessimally small
RR relatively small

They move on.

RR’s WS2 Par 24 "The process that I describe at paragraph 25 above could be carried out without the
Subpostmaster’s knowledge (which Dr. Worden acknowledges in principle at paragraph 1134)
and in my recollection it …"

“… sometimes was done without the Subpostmaster’s prior knowledge,
for example if the Subpostmaster was away from the branch on a lunch break and had not
logged out of the system. Whilst in some circumstances we did notify the Subpostmaster…"

" in
advance in order that they did not use the counter being worked on, there were times where
the job needed to be done quickly to prevent potentially catastrophic failure and we were
unable to contact the Subpostmaster beforehand."

POQC any examples?
RR no
POQC if you cou;dn’t contact the SPM before would you make sure they knew
RR afterwards, yes

POQC concludes xe of RR.

Patrick Green QC on his feet for the claimants. He is taking RR to a document,

PG you’ve got closure category codes and the legend or description against them… “program approved no fix required” lots of them. What does that mean?

RR no idea
PG “administrative response” what does that mean?
RR don’t know
PG there are more than 5000 of them- now the court was shown a document which you haven’t seen...

… which explains the codes.

63 - programme approved no fix required - “there IS a fault in the product” at this distance in time can you remember any examples of this

PG Describing PEAK 2207887 in which a software error created a £1,085,000 discrepancy.
Did that ever cross your desk?
RR no
PG you were asked about pressure in relation to SLAs

RR told by J he can leave the witness box and “you can chat to people about the case if you’d like to"

Ian Henderson director of Second Sight in the witness box. He has told the court he won’t give opinions and that he cannot speak about some areas because of his contractual agreement with the Post Office.

He is not going to be quizzed about any of the opinions in Second Sight’s report. This will all be about findings of fact.

IH we were first appointed July 2012 and towards the end of 2012 we were in active discussion with JFSA and PO. After we issued our interim report...

… PO decided we should move towards the mediation scheme.

POB - there were 150 mediation scheme applicants
IH - and 136 were accepted onto the scheme
POB you finished my question.

POB let’s have a look at the 2S engagement letter - your role in the scheme
IH letter is dated July 2014 - a long time after we started work on the mediation scheme.
POB well I want to look at this. If you have concerns you can raise them.

4.2 services being supplied by 2S directors - that’s you and one other
Here are the scope of services - ie what you have to do.
IH I’ve read that
POB does that accord with your recollection of what your role was
IH this letter arrived half way through the mediation scheme

IH we’d already done a great deal of work possibly on different terms described in this letter
POB it was recognised in an early stage that 2S might not be qualified to adjudicate on every issue
IH I take issue with “early stage” this letter is not written at an early stage...

… of the scheme.
POB but thereafter it was recognised?
IH yes

IH point 5.1 in this letter was a degree of variation on our activities
POB but you agreed it
IH half way through the mediation scheme, yes.

Onto briefing report in July 2013.

POB Part 1 report was intended to be a largely uncontroversial?
IH yes uncontroversial to reflect the permanent relvant information. Part 2 was the thematic issues.

POB there were two part 2 reports?
IH yes both of which had substantive responses from PO to them
Now onto key documents used in the mediation scheme.
POB without apportioning blame to anyone would it be fair to say the mediation scheme

went a lot more slowly than anyone expected.
IH agrees
POB reads from letter from Sir Anthony Hooper to an MP about the slow pace of the mediation scheme from Dec 2014

POB do you recall seeing this update on progress
IH I don’t remember this letter or appendix, but it does look broadly familiar so we may have discussed it at the Working Group
[boy this talk of CQRs CRRs, CEDR and Working Groups is taking me back. Heady days. Joke.]

IH makes a general point on slow progress of mediation scheme - it took some time for CEDR to be appointed and I think that more than anything else reflects the relatively small number of cases on the schedule.

POB goes to CEDR doc. which report 31 cases for mediation of which 12 had been readied.
So 18 months after scheme closed there had only been 12 mediations - is that right?
IH this letter is to PO GC so I wouldn’t have seen it, but the progress timeline is about right.

IH saying by the time 2S were fired they’d completed their work on 116 of the accepted applicants so they were nearly done.
POB showing the court their termination letter and a longer letter in which PO sets out a plan for how 2S can finish its outstanding work.

POB the letter makes clear that there could be a working future for PO and 2S.
IH you haven’t mentioned a 3rd document which was a press release announcing the termination of the mediation scheme which I think is the primary document

POB but the mediation scheme continued didn’t it
IH yes but the working group was terminated the day before it was due to meet which was a considerable shock to all concerned.

POB but your work on v2 of the briefing report part 2 continued
IH and we were invited to complete it by early april so we continued to work on that.
POB there is then an exchange of letters in which the terms of 2S and PO ongoing relationship was agreed.

IH one element we found unacceptable was that we would be instructed directly by claimants. We said that could compromise our independence and refused and PO accepted that.

Cross examination finishes.

J has some questions. You don’t need to tell me who told you, but you were told your first WS was too long
IH I was told it stood the risk by the court of being consdered too long and i should reduce

J alright - but no such restriction was made by this court.
J and you were told you couldn’t discuss certain things because of your contractual relationship with the PO
IH yes it has been modified by the PO to allow us to appear in court under a protocol
J but it has narrowed...

… the scope of the evidence you are able to give to the court.
IH correct
J takes him to the 2014 engagement letter - clause at 6.2 in that letter - 2S will not act directly or indirectly in any capacity against the Post Office save to the extent any work...

… done will not undermine.
Was that clause in the initial contract?
IH I don’t recall?
J was this letter a change in terms
IH we anticipated we might be asked to give factual evidence or support in some way to claimants in a legal action. We were quite up front about it...

… and i remember a 12 month period when we were barred from acting from the claimants and I agreed on a handwritten note which extended that to 15 months and I note that’s in the letter here.
POB back on his feet. Did you feel inhibited in answering any questions I asked?

IH I had the protocol agreement at the back of my mind as I answered those questions
POB but there was no risk of that in the questions I asked
IH I’d have to refresh my memory of the questions. I’m sorry.
J I think, Mr Draper [POB] that’s rather something for me to deal with.

IH has finished his very brief cross examination.

PG QC now on his feet asking for 3rd party disclosure order to get some documents from Royal Mail.
J asks PO QC who acknowledges he has no brief but recalls being told by RM they would not surrender docs without an order.

J and you need this for Monday? I can’t make an order to the RM for Monday
PG brings up a document noting concern from RM rather than a refusal.
J says its’ a logistical problem I can’t make an order in these circs about that. I can express a view on the transcript...

… and I will make an order. RM are not here today so I can’t make a 3rd party disclosure but I will order the claimants to issue an application for the Ernst and Young report by 1pm tomorrow and it can be issued to RM by Monday. If they want to make representations...

… they can come at 10am on Tuesday.

Hopefully by reading the transcript which I am sure you will show them they will see there is very little to be gained by being uncooperative. But I will not pre-judge the order.

Judge now wants news on redaction of 31 documents. POQC says
16 documents remained redacted on privilege
4 were confidential or irrelecant
7 docs had redactions removed or reduced because they had been applied too widely. POQC shows J an example...

In relation to redactions they were narrowed or removed from 4 docs for pragmatic reasons.
J confirms these numbers.
POQC gives another example to satisfy his lordship...

… these were redacted on grounds of relevance.
J well redaction on relevance is potentially a two-edged sword.

[his lordship seems satisfied by the example]
[we move on]

Closing arguments dates for this trial are going to shift again. Now looking like 7th and 8th May not 8th and 9th.

J about to spell out what is going to happen wrt to tomorrow’s judgement.

J I am going to hand it down at 12 o’clock tomorrow.

It will be open to the public - there will be paper copies but it will be instantly posted on the relevant websites

That’s it. We are done. Back in court 12 noon tomorrow for the judgment.

Thanks for reading.

I’ll get everything posted up asap to postofficetrial.com - especially the transcript as there was some good stuff in there this afternoon…

@threadreaderapp unroll pls

Day 4 transcript here: postofficetrial.com/2019/03/horizo…

Ian Henderson witness statement here: postofficetrial.com/2019/03/horizo…

#postofficetrial