<Muttley@dastardlyhq.com> wrote:
> On Mon, 4 Jul 2022 13:50:07 +0100
> Roland Perry <roland@perry.co.uk> wrote:
>> In message <03m5chp8csd24tarcv38fcr23d1p1pgssl@4ax.com>, at 13:14:23 on
>> Mon, 4 Jul 2022, Recliner <recliner.usenet@gmail.com> remarked:
>>
>>> We all know that the Elizabeth Line has been plagued with software
>>> problems, but I didn't expect, as a passenger, to be
>>> able to spot a very visible bugs= more than a month after opening.
>>
>>> At Whitechapel station last week, I spotted that the displays over the
>>> PEDs seem not to know that the last couple of trains had arrived and
>>> left, and still showed them as 'Due'. So they showed several trains all
>>> due at once, when the trains had actually arrived on time, and left.
>>> The trains were running punctually, without problems, but the PED PISs
>>> seem to have got stuck in a loop
>>>
>>> <https://www.flickr.com/photos/recliner/albums/72177720300301029>
>>
>> I didn't notice that, but at the one place I had time to observe, the
>> x-minutes on the doors was consistently one minute more optimistic than
>> the "secret" display intended for platform staff on the back wall of the
>> platform.
>>
>> Trains being self-evidently not-arrived when the PIS said they were, or
>> the PIS having them in an impossible future arrival order, is something
>> I've seen quite a few times on National Rail stations.
>>
>> I've attributed the impossible-arrival-order to perhaps the PIS at
>> multi-TOC stations being fed separately by the different TOCs. And as
>> one train can't overtake another [at such station approaches] and hence
>> destined to arrive one after the other, the various timings therefore
>> contradicting one another because the train-behind was temporarily
>> showing as arriving before the one ahead of it.
>
> The sane approach would have them fed direct from the signalling system but
> I guess thats too much of an ask. As long as it looks flash who cares if its
> particularly accurate I imagine is their thinking.
>
>

It is fed from the signalling system (well, from the train running data
which is updated from the signalling system), and it doesn't look 'flash'.
It's just not clever enough to understand the track layout, so a faster
train will still show on time until it actually gets delayed at a timing
point, by the slower train which it can't overtake.

Anna Noyd-Dryver

On 04/07/2022 18:38, Anna Noyd-Dryver wrote:
> <martin.coffee@round-midnight.org.uk> wrote:
>> On 04/07/2022 16:41, Anna Noyd-Dryver wrote:
>>> Roland Perry <roland@perry.co.uk> wrote:
>>>> In message <03m5chp8csd24tarcv38fcr23d1p1pgssl@4ax.com>, at 13:14:23 on
>>>> Mon, 4 Jul 2022, Recliner <recliner.usenet@gmail.com> remarked:
>>>>
>>>>> We all know that the Elizabeth Line has been plagued with software
>>>>> problems, but I didn't expect, as a passenger, to be
>>>>> able to spot a very visible bugs= more than a month after opening.
>>>>
>>>>> At Whitechapel station last week, I spotted that the displays over the
>>>>> PEDs seem not to know that the last couple of trains had arrived and
>>>>> left, and still showed them as 'Due'. So they showed several trains all
>>>>> due at once, when the trains had actually arrived on time, and left.
>>>>> The trains were running punctually, without problems, but the PED PISs
>>>>> seem to have got stuck in a loop
>>>>>
>>>>> <https://www.flickr.com/photos/recliner/albums/72177720300301029>
>>>>
>>>> I didn't notice that, but at the one place I had time to observe, the
>>>> x-minutes on the doors was consistently one minute more optimistic than
>>>> the "secret" display intended for platform staff on the back wall of the
>>>> platform.
>>>>
>>>> Trains being self-evidently not-arrived when the PIS said they were,
>>>>
>>>
>>> Did they arrive very shortly afterwards?
>>>
>>> If so, it seems likely that they change to 'arrived' when they enter th
>>> track circuit in which the platform is situated. In certain locations this
>>> might obviously be some distance before the platform itself. An obvious fix
>>> would be to add a fixed delay (location-specific) between the track circuit
>>> activation and the display changing.
>>>
>>>> or
>>>> the PIS having them in an impossible future arrival order, is something
>>>> I've seen quite a few times on National Rail stations.
>>>>
>>>> I've attributed the impossible-arrival-order to perhaps the PIS at
>>>> multi-TOC stations being fed separately by the different TOCs.
>>>>
>>>
>>> No, it'll be fed from the same NR database as RTT, Traksy etc.
>>>
>>>> And as
>>>> one train can't overtake another [at such station approaches] and hence
>>>> destined to arrive one after the other, the various timings therefore
>>>> contradicting one another because the train-behind was temporarily
>>>> showing as arriving before the one ahead of it.
>>>
>>> I can think of two reasons for this to happen. Most obvious is a converging
>>> junction just before the station, and the signalling decision hasn't get
>>> been made.
>>>
>>> More common, perhaps, is a situation where a stopping train is running
>>> late, and being caught up by a faster train. The PIS system just isn't
>>> intelligent enough to take account of the track layout etc., so it gives
>>> the faster train it's booked arrival time until it actually catches up with
>>> the slower train and gets delayed passing one of the timing points.
>>
>> The departure boards at Swindon railway station can be quite exciting
>> when all services are delayed. Not only do the arrival times
>> continuously change but so does the platform.
>>
>
> Is that from recent (post-TVSC) experience, or pre-resignalling?

It was during one of the electrification closures when the Cardiff /
Portsmouth services were diverted to Swindon which was very satisfactory
as the Paddington / Paddington services were diverted via Gloucester.

I cannot remember the date.

>
> IMX Swindon Panel (two signallers in a 1960s building behind platform 4,
> using 1960s tech) used to be very adept at platform swaps between 1 and 3
> (ie opposite sides of the island platform, so minimal impact on passengers)
> to keep trains flowing as freely as possible.
>
> Since transfer to TVSC (one signaller now responsible for the same area,
> using a computer screen in a building at the far end of Didcot Parkway car
> park), much more reliance is placed on ARS (Automatic Route Setting)
> (because the one signaller now controls a larger area), which tends to
> result in trains running through their booked platform no matter what -
> even if that means other trains held outside the station while there's an
> empty platform they could be using.
>
>
> Anna Noyd-Dryver
>

Roland Perry <roland@perry.co.uk> wrote:
> In message <t9v2a6$3dpop$1@dont-email.me>, at 15:51:35 on Mon, 4 Jul
> 2022, Anna Noyd-Dryver <anna@noyd-dryver.com> remarked:
>> Roland Perry <roland@perry.co.uk> wrote:
>>> In message <t9urr6$3d3gl$1@dont-email.me>, at 15:01:09 on Mon, 4 Jul
>>> 2022, Certes <none@nowhere.net> remarked:
>>>> On 04/07/2022 13:14, Recliner wrote:
>>>>> We all know that the Elizabeth Line has been plagued with software
>>>>> problems, but I didn't expect, as a passenger, to be
>>>>> able to spot a very visible bugs= more than a month after opening.
>>>>> At Whitechapel station last week, I spotted that the displays over
>>>>> the PEDs seem not to know that the last couple of
>>>>> trains had arrived and left, and still showed them as 'Due'. So they
>>>>> showed several trains all due at once, when the
>>>>> trains had actually arrived on time, and left. The trains were
>>>>> running punctually, without problems, but the PED PISs
>>>>> seem to have got stuck in a loop
>>>
>>>>> <https://www.flickr.com/photos/recliner/albums/72177720300301029>
>>>>>
>>>> It sounds as if a bit of kit that should detect when a train reaches
>>>> that platform, and is thus no longer Due, doesn't.
>>>
>>> Wouldn't that be the mission critical bit of kit which says "Ok, train
>>> has reached the platform, it's now OK to open the PEDs"?
>>>
>>> Speaking of which, something I don't think I've seen mentioned before is
>>> that with the trains shorter than [all/most?] platforms at the moment,
>>
>> "At the moment" meaning "for the foreseeable future", presumably, as I'm
>> not aware of any expectation having been set for extending class 345 beyond
>> 9 cars? (As opposed to 'making allowance for this to happen when required
>> in the future').
>>
>>> do they stop in the middle, or at one end.
>>
>> From observation, it's at the end; the arriving train draws right to the
>> far end of the platform.
>
> I'd need to do another field trip to confirm that. In particular
> eastbound trains at Farringdon where they clearly expect most pax to
> exit towards the rear.
>

From examination, I'm going to tentatively change my answer to 'middle'. At
the Liverpool Street end of Liverpool Street station, there's a length of
platform with no doors (you have to turn back from the entrance from the
escalators to get to it, but I think the lift brings you down right at the
end). Examination of the Eastbound platform revealed a similar length of
no-doors. Looking out from the departing train, I judge there to be a
similar length at the other end.

At Farringdon there's a similar length at the Barbican end, though it was
difficult to judge from a moving train whether the length at the Farringdon
end was a similar size.

>>> And if at one end, who chooses which end,

The designers of the station, prior to installation of signalling and PEDs.

>>> and do the relevant PIS
>>> over the never-going-to-open PEDs tell passengers to wait elsewhere?
>>
>> I'm not sure OTTOMH (and can't check for a couple of hours) that there are
>> even PEDs fitted in the platform screen there; OTOH having seen someone who
>> wasn't paying attention waiting there and having to trot down to the rear
>> of the train when it arrived, maybe there are.
>>

As expected, it's a glass screen with no doors or recesses for future doors
[1], and thus no indication above the not-doors to say that they are not
doors.

<https://twitter.com/annanoyddryver/status/1544031609774538753>

[1] though the sizes of the glass panels appear to match the door size and
spacing of another carriage.

Anna Noyd-Dryver

Anna Noyd-Dryver <anna@noyd-dryver.com> wrote:
> Recliner <recliner.usenet@gmail.com> wrote:
>> Roland Perry <roland@perry.co.uk> wrote:
>>> In message <2p06cht77kmt1jfq7jc0iqlcg0c335ud64@4ax.com>, at 16:18:32 on
>>> Mon, 4 Jul 2022, Recliner <recliner.usenet@gmail.com> remarked:
>>>> On Mon, 4 Jul 2022 15:29:13 +0100, Roland Perry <roland@perry.co.uk> wrote:
>>>>
>>>>> In message <t9urr6$3d3gl$1@dont-email.me>, at 15:01:09 on Mon, 4 Jul
>>>>> 2022, Certes <none@nowhere.net> remarked:
>>>>>> On 04/07/2022 13:14, Recliner wrote:
>>>>>>> We all know that the Elizabeth Line has been plagued with software
>>>>>>> problems, but I didn't expect, as a passenger, to be
>>>>>>> able to spot a very visible bugs= more than a month after opening.
>>>>>>> At Whitechapel station last week, I spotted that the displays over
>>>>>>> the PEDs seem not to know that the last couple of
>>>>>>> trains had arrived and left, and still showed them as 'Due'. So they
>>>>>>> showed several trains all due at once, when the
>>>>>>> trains had actually arrived on time, and left. The trains were
>>>>>>> running punctually, without problems, but the PED PISs
>>>>>>> seem to have got stuck in a loop
>>>>>
>>>>>>> <https://www.flickr.com/photos/recliner/albums/72177720300301029>
>>>>>>>
>>>>>> It sounds as if a bit of kit that should detect when a train reaches
>>>>>> that platform, and is thus no longer Due, doesn't.
>>>>>
>>>>> Wouldn't that be the mission critical bit of kit which says "Ok, train
>>>>> has reached the platform, it's now OK to open the PEDs"?
>>>>>
>>>>> Speaking of which, something I don't think I've seen mentioned before is
>>>>> that with the trains shorter than [all/most?] platforms at the moment,
>>>>
>>>> In the core, yes, but outside the core, many platforms are shorter
>>>> than the trains, which warn of selective door opening at some stations.
>>>>
>>>>> do they stop in the middle, or at one end. And if at one end, who
>>>>> chooses which end, and do the relevant PIS over the never-going-to-open
>>>>> PEDs tell passengers to wait elsewhere?
>>>>
>>>> I don't think there are any PEDs currently installed for
>>>> longer-than-nine carriage trains. So, all the installed PEDs will open.
>>>
>>> There are, because I saw them. Perhaps at the eastern end of Farringdon?
>>>
>>> If not never-opening-PEDs what other thing might they have between the
>>> platform and the train, when the train is shorter than the platform?
>>
>> Just a fixed screen that could later be replaced by PEDs. So there would be
>> the anchor points for the PEDs, and maybe even the wiring. There might also
>> be emergency tunnel access/escape doors, normally kept locked from the
>> platform side.
>>
>>
>>>> The core platforms are long enough to allow longer trains in the
>>>> future, but none are currently planned. If such trains are introduced,
>>>> I think there's passive provision for more PEDs to be installed.
>>>>
>>>>> Another potential failure, is when I got off at Paddington westbound,
>>>>> which has numerous on-platform and in-train announcements about everyone
>>>>> getting off, a late-arriving couple dashed onto the platform [inevitably
>>>>> wrong one at this stage] and onto the train, then the doors shut.
>>>>
>>>> It sounds like there's no walk through by staff? If no-one stops
>>>> passengers travelling to the turnback sidings, I'd have thought every
>>>> enthusiast would be trying to get a ride there.
>>>
>>> I didn't wait round to see what happened, but a walk-through is going to
>>> take several minutes. And there's another train arriving in 5-7 mins.
>>
>> Less than that. The trains are every five minutes, so by the time the doors
>> have opened and the passengers alighted, the next train is probably only
>> four minutes away. As the trains are very open, and platform straight,
>> maybe a couple of members of platform staff just take a quick look in from,
>> say, doors in car 3 and car 7?
>>
>
> The high-backed seating bays in every carriage except 1 and 9 would prevent
> that being effective.
>

Observation says that three or four platform staff check the train,
checking two-three coaches each. They walk through and then press the door
close button as they alight, which closes the train doors, but not PEDs, on
'their' 2-3 carriages. Once they've all alighted, the PEDs are closed and
the train departs, but not immediately. I was at the wrong end to check
what the driver does re. turnback.

Anna Noyd-Dryver

On Mon, 4 Jul 2022 17:38:44 -0000 (UTC)
Anna Noyd-Dryver <anna@noyd-dryver.com> wrote:
><Muttley@dastardlyhq.com> wrote:
>> The sane approach would have them fed direct from the signalling system but
>> I guess thats too much of an ask. As long as it looks flash who cares if its
>> particularly accurate I imagine is their thinking.
>>
>>
>
>It is fed from the signalling system (well, from the train running data
>which is updated from the signalling system),

Bit worrying in that case that it can't even get train orders correct.

>and it doesn't look 'flash'.

It is flash. You don't need a full colour display with a partial line
diagram and "Boarding" messages (no! really? Is that why the doors are open?)
and lots of scrolling waffle just to indicate the next train. A cheapo dot
matrix as use elsewhere would do just as well and save a lot of cash.

>It's just not clever enough to understand the track layout, so a faster
>train will still show on time until it actually gets delayed at a timing
>point, by the slower train which it can't overtake.

That's no excuse. It should know the train order and its trivial to code
in the fact that trains can't overtake each other (on this line anyway) and
adjust arrival times accordingly.

On 04/07/2022 13:50, Roland Perry wrote:
> I've attributed the impossible-arrival-order to perhaps the PIS at multi-TOC stations being fed separately by the different TOCs. And as one train can't overtake another [at such station approaches] and hence destined to arrive one after the other, the various timings therefore contradicting one another because the train-behind was temporarily showing as arriving before the one ahead of it.

I've seen it several times when waiting on the northbound platform at Blackfriars. When trains are due from London Bridge and from Elephant & Castle and one is a few minutes late, sometimes the same arrival time is shown on the synoptic display, and on the platform the order of 1st and 2nd trains sometimes switches a couple of timed before the situation resolves when one of them actually turns up.

But this is not because more than one TOC is involved, these are all Thameslink services.

I simply think it is lazy programming - the programmer concerned didn't realise that showing two trains due at the same platform at the same time is pretty stupid. And also having the order switch a few times is also not helpful to passengers: I don't think the points at the junction actually switch first one way and then the other, it seems more likely that the information internally has predicted arrivals given to the second, and these are continually fluctuating.

The signalling system surely knows which way the junction just south of Blackfriars is going to be set and should be able to inform the passenger information system accordingly.

--
Clive Page

Clive Page <usenet@page2.eu> wrote:
> On 04/07/2022 13:50, Roland Perry wrote:
>> I've attributed the impossible-arrival-order to perhaps the PIS at
>> multi-TOC stations being fed separately by the different TOCs. And as
>> one train can't overtake another [at such station approaches] and hence
>> destined to arrive one after the other, the various timings therefore
>> contradicting one another because the train-behind was temporarily
>> showing as arriving before the one ahead of it.
>
> I've seen it several times when waiting on the northbound platform at
> Blackfriars. When trains are due from London Bridge and from Elephant &
> Castle and one is a few minutes late, sometimes the same arrival time is
> shown on the synoptic display, and on the platform the order of 1st and
> 2nd trains sometimes switches a couple of timed before the situation
> resolves when one of them actually turns up.
>
> But this is not because more than one TOC is involved, these are all Thameslink services.
>
> I simply think it is lazy programming - the programmer concerned didn't
> realise that showing two trains due at the same platform at the same time
> is pretty stupid. And also having the order switch a few times is also
> not helpful to passengers: I don't think the points at the junction
> actually switch first one way and then the other, it seems more likely
> that the information internally has predicted arrivals given to the
> second, and these are continually fluctuating.
>
> The signalling system surely knows which way the junction just south of
> Blackfriars is going to be set and should be able to inform the passenger
> information system accordingly.
>

If it's a converging junction, until the route is actually set across the
junction, either train could be arriving first.

I do agree that the system should only change once, though, and that be
when the route is actually set.

Anna Noyd-Dryver

On Tue, 5 Jul 2022 10:53:51 +0100, Clive Page <usenet@page2.eu> wrote:

>On 04/07/2022 13:50, Roland Perry wrote:
>> I've attributed the impossible-arrival-order to perhaps the PIS at multi-TOC stations being fed separately by the different TOCs. And as one train can't overtake another [at such station approaches] and hence destined to arrive one after the other, the various timings therefore contradicting one another because the train-behind was temporarily showing as arriving before the one ahead of it.
>
>I've seen it several times when waiting on the northbound platform at Blackfriars. When trains are due from London Bridge and from Elephant & Castle and one is a few minutes late, sometimes the same arrival time is shown on the synoptic display, and on the platform the order of 1st and 2nd trains sometimes switches a couple of timed before the situation resolves when one of them actually turns up.
>
I've seen later trains shown arriving first at Clapham Junction with
London-bound SW trains. IIRC the sequence seems to go -
1.Trains arriving in sensible order matching times.
2.Times change with trains now in illogical order.
3.Order changes.

The difference elsewhere (e.g. Up trains at Harrow) is that No.2 does
not occur so the Clapham version seems to be "thinking out loud" while
shuffling the information.
<snip>