Mastodawn

2 weeks ago, girlfriend and me took a #nighttrain from Berlin Südkreuz to Budapest (EN 40457). On paper a 12 hour ride (20:25 - 08:29), in reality we arrived after 10 in the morning, since the train just stood around in Bratislava for at least an hour ("something with the rails" according to the conductor) and we started out with a >30minutes delay from Berlin.

I requested compensation through the #ÖBB -bot (which worked very well actually) and got a reply later on: denied, system says no, you arrived with less than an hour of delay.

What I noticed during our ride, is that somewhere after Břeclav (Czechia), our train state in the DB app stopped updating and it didn't take our Bratislava-delay into account anymore.

#crossBorderRail

What I keep wondering though, why did the train schedule didn't update anymore..

Is there a technical reason that these systems aren't linked in Slovenia/Hungary? A checkbox that has not been set? On purpose for annoying people requesting refunds like me?

#crossBorderRail #bahnBubble #öbb

By the way, after another mail from my side explaining the situation the refund was accepted a day later, so this is not meant as a complaint to ÖBB or their process.

In my mail I mentioned that I also saw the planning wasn't updated from (at least) Bratislava and that indeed the ÖBB-system probably showed we arrived with a 30-45 minutes delay, not a >1 hour delay.
I also asked how I could in this case show that we were indeed delayed by over an hour. ÖBB advises to make a screenshot of the train schedule up to 1 week after traveling, but that also showed the wrong times.

Oliver Blanthorn 3d ago

@quin_antarctic personally I take a photo of my watch or arrivals display with the train in the background as a kind of insurance policy :)

Cycling on Rails 3d ago

@bovine3dom @quin_antarctic That's assuming the railway still has an arrivals board. SBB, an otherwise respectable company, is particularly evil in that regard: on-platform displays only show departures or a "please don't enter" placeholder at the final stop AND their app only shows forward-looking delays but instantly forgets any actual delay of the previous stops. If you're past the theoretical arrival time of the last stop the train may be completely erased from the app.

波鉄 (Hatetsu)3d ago

@cycling_on_rails @bovine3dom @quin_antarctic They'd probably argue it's less confusing, but I agree - displays should show where trains are coming from, too. And, ironically, where they do have arrival boards, they do better than most everyone else by making them clearly contrast departure boards with an all-white aesthetic versus blue screens with yellow borders (and generally being one screen next to two or three screens' worth of departures). The SNCF convention of blue for departures and green for arrivals that's slowly spreading to Poland for some reason (more proof of my theories that Polish and French railway history are intertwined in mysterious ways) seems not very colorblind-friendly, and with non-LCD boards they often look nearly identical at a glance

Oliver Blanthorn 3d ago

@HaTetsu blue/green seems to be the standard in lots of countries. definitely austria, maybe czech republic? maybe Hungary?

@cycling_on_rails

Cycling on Rails 3d ago

@quin_antarctic I also often notice the same thing where delays are not properly tracked across borders. Train operators generally only see delay within their "home" country but default to "on time" once abroad.

In my opinion, this is why the upcoming EU regulation on passenger rights should include mandatory tracking of delays across Europe in an open data manner. There cannot be effective compensation if operators pretend delays don't happen. cc @erpu @jon @stefanlindbohm @bovine3dom

Cycling on Rails 3d ago

@quin_antarctic @erpu See my experience last December on the Munich-Warsaw night train, where finding an accurate source for the delay was challenging: https://mastodon.social/@cycling_on_rails/115807268291974432

Basically, if the Interrail app cannot get the delay data right, there's a serious open data problem from the railway operator.

@cycling_on_rails @erpu Ah wait, I forgot about the part where this train was indeed also split up/combined. Our train was a combined EN40457/NJ457 combination from Berlin (from Warsaw actually I guess): EN to Budapest, NJ to Vienna.
So I assume ÖBB/DB will have ÖBB-data (Nightjet) up until Břeclav. Then it is split up/combined and I guess they are dependent on the Hungarian MAV for info (that should be the operator on the EN40457). Which didn't work then.

Cycling on Rails 3d ago

@quin_antarctic There are more details here: https://www.vagonweb.cz/razeni/vlak.php?zeme=DB&kategorie=EN&cislo=40457/@&nazev=&rok=2026&lang=en

Essentially yes, the Břeclav-Budapest segment combines trains coming from Berlin, Prague, Warsaw & Przemyśl. I would say that the Hungarian railways MÁV are most likely to have had the delay data upon arrival?

vagonWEB » Train compositions » 2026 » DB EN » EN 40457

@cycling_on_rails Hmm, looking at MÁV's site (https://mavplusz.hu), the timetable for this connection of last week shows not a single delay, and it doesn't go back further than last Sunday.
The planner doesn't go back at all.
So maybe it did, but definitely not anymore.

Come to think of it: we booked our tickets through ÖBB, you'd also expect the ÖBB to have a requirement in place, that they are able to see delays of trains for which they sell tickets, to prevent this kind of thing in the first place.

Menetrend és térkép: vonat, VOLÁN busz, helyi járat | MÁVPlusz

Menetrendek és térkép: valós idejű vonat, VOLÁN busz, HÉV és TramTrain menetrend. Útvonaltervező helyközi és helyi utazáshoz, akár Budapesten is. | MÁVPlusz

@cycling_on_rails I didn't know about vagonweb.cz yet though, another black hole to lose myself in I guess... Thanks!

Cycling on Rails 3d ago

@quin_antarctic Yeah, it's a pretty cool resource! I really like the photo gallery, probably the most complete "Wikimedia for train pictures" out there 🙃 I'm thinking to contribute once I get the time to sort my pictures as there are still occasional gaps here and there.

Oliver Blanthorn 3d ago

@cycling_on_rails IIRC the new TSI regs do have some clauses trying to prevent such orphaned trains. so in theory it should get better within the next five years

(and delay data must be released with a real open licence as opposed to the highly commercially sensitive stuff like "where is the Speisewagen?")

@wnd might know more

@quin_antarctic @erpu @jon @stefanlindbohm

Dr. Sobek 3d ago

@bovine3dom @cycling_on_rails @wnd @quin_antarctic @erpu @jon @stefanlindbohm What is a Speisewagen ?

Cycling on Rails 3d ago

@Sobex @bovine3dom Restaurant car in German 🙃

Dr. Sobek 3d ago

@cycling_on_rails How is that commercially sensitive ?

Cycling on Rails 3d ago

@Sobex It definitely shouldn't be. I think @bovine3dom might refer to the fact that GTFS open data is lacking such detailed information about train composition, but you can purchase that information for a hefty fee from the railway operator or a paid database like MERITS (about 50k€/year)?

Or some operators (SNCF?) simply only distribute such information in their own ticketing app behind a login or captcha, preventing 3rd-parties from displaying useful information to passengers?

Oliver Blanthorn 3d ago

@cycling_on_rails close! you gotta swot up more on the TSI regs

Oliver Blanthorn 3d ago

@Sobex sorry it was a joke for maybe three people in the world

the new TSI regs force railways to publish train formation (which carriage is where) but only requires them to release it under "no derivatives, non-commercial" usage restrictions

which is clearly completely bonkers because what kind of unfair commercial advantage can you get from knowing where first class is? or the restaurant car?

@cycling_on_rails

Dr. Sobek 3d ago

@bovine3dom @cycling_on_rails Hmm, so what is considered Derivative work in this case.
(I assume the publication is done as a textual format, not a nice picture ? Right ?)

Looking up an image corresponding to the train and displaying that would be derivative ?

Oliver Blanthorn 3d ago

@Sobex honestly, no idea. i think the main thing it tells you is that the person/committee who picked that licence has never built anything in their life :)

@cycling_on_rails

Dr. Sobek 3d ago

@bovine3dom @cycling_on_rails Yeah ND is kind of only suitable for a finished product that can be redistributed as is. 🤦‍♂️

Oliver Blanthorn 3d ago

@Sobex not a lawyer, don't want to be a lawyer

if you've picked a licence where everyone who might actually use the data says "what the fuck does that even mean?" then i would argue that you have picked a bad licence

@cycling_on_rails

Will Deakin 3d ago

@Sobex @cycling_on_rails that would be a question for the railway undertaking.

Noting that on the GB network, which has a relatively permissive open data culture, it was only agreed to publish train consist information last year.

Dr. Sobek 3d ago

@wnd @cycling_on_rails
Edit: This was not an objection in the UK.

Well, train consist info is the sort of thing that can change unexpectedly last minute. ( @hopla how often do you end up with a different rolling stock as planned on your trains).

I've had a few instances of rolling stock changes on mandatory reservation trains (which leads to massive mess for plenty of reasons).

Will Deakin 3d ago

@Sobex @cycling_on_rails @hopla fair but that was not the objection. It was based on a concern about commercial advantage to competitors.

The vast majority of the time the consist is as planned and, if changed, this data is known and shared.

For example, a driver has know how to safely operate a train, a signaller has to know whether an alternative path through interlocking during perturbation is safe.

It's up to you if don't want to know, but don't make up reasons for those that do.

Dr. Sobek 3d ago

@wnd @cycling_on_rails @hopla I’m editing my message to reflect that my hypothesis was wrong then.

Will Deakin 3d ago

@Sobex @cycling_on_rails @hopla it is a reasonable statement though. RUs and IMs are very bad at sharing this information with passengers

Dr. Sobek 3d ago

@wnd @cycling_on_rails @hopla I do see this can be an issue, as a wrong information can be worse than no information, and this is typically a case where stale data can more likely remain.

(Cache invalidation is one of the two great problem in CS)

Oliver Blanthorn 3d ago

@Sobex can't tell you that, that's commercially sensitive. we'd have to negotiate terms first otherwise it's CC-NC-ND

@cycling_on_rails @wnd

Dr. Sobek 3d ago

@cycling_on_rails @quin_antarctic @erpu @jon @stefanlindbohm @bovine3dom Perhaps mandate that each national rail network operator provides a tracking interface (at least via train number + date), in which case, you should be able to get accurate delay on arrival from the local rail network operator.

What do you think ?

Will Deakin 3d ago

@cycling_on_rails @quin_antarctic @erpu @jon @stefanlindbohm @bovine3dom
as I understand it, the February 2026 publication of Telematics TSI does not mandate the tracking of delay information, but rather the open publication of all train service tracking and timetable data.
Whether a train service is delayed, or heavens forfend, on-time, is the left as an exercise for the reader.
I also note that, although this is now in law, there is still an implementation period which has yet to elapse.

@wnd @cycling_on_rails @erpu @jon @stefanlindbohm @bovine3dom I can imagine that you don't want to track delay information as a separate field per se, but I'd assume you do want to be able to track the actual train arrival/departure times. Or is that what's included with "all train service tracking and timetable data"?

I'd assume the planned arrival/departures are known, so any differences are pretty easily calculated.

Also seems very relevant information anyhow and I assume (again), that this information is available to the operators/planners anyway..

Will Deakin 3d ago

@quin_antarctic @cycling_on_rails @erpu @jon @stefanlindbohm @bovine3dom yes, this is what I meant by "all train service tracking and timetable data".

Determining whether or the cause of a delay or otherwise is more a performance analysis activity. About which the TEL TSI has very little to say.

Stefan Lindbohm 3d ago

@wnd My interpretation in practical terms is that all the (raw) data should by law be available by 2029, but law says nothing about who needs to build a UI to show it.

Realtime data is by definition split at borders (infrastructure managers change + different NAP’s for the data), so there’s nothing saying any particular app would add these data sources just because they exist.

@quin_antarctic @cycling_on_rails @erpu @jon @bovine3dom

Stefan Lindbohm 3d ago

@wnd However, there is a UIC-led project for state railways (mostly) to share data in a proprietary manner ahead of the legislated stuff. So certain countries’ state railways should get better at this stuff soon.

@quin_antarctic @cycling_on_rails @erpu @jon @bovine3dom

Cycling on Rails 3d ago

@stefanlindbohm @wnd @quin_antarctic @erpu @jon @bovine3dom Ah, looks like UIC needs to find a way to justify their hefty MERITS fee ahead of mandatory open data! Which confirms that yes, mandatory open data is a good incentive to kick their b*tts and improve the data sharing. 👍

Will Deakin 3d ago

@stefanlindbohm @quin_antarctic @cycling_on_rails @erpu @jon @bovine3dom there is also a discussion with ERA and RNE to talk about making this information, which RNE mostly has open. But I don't know where that has got to yet.

Oliver Blanthorn 3d ago

@wnd the older crappier passenger rights regs say if you're selling tickets you have to display realtime data if it's available. and tsi makes it available...

@stefanlindbohm @quin_antarctic @cycling_on_rails @erpu @jon

Will Deakin 3d ago

@stefanlindbohm @quin_antarctic @cycling_on_rails @erpu @jon @bovine3dom I agree with the UI* stuff. But, and this is important, the linkage between paths and services is maintained across borders, and that this linkage is data is held in each message sent.

1/n
* but who needs a UI, they are difficult to build** and you can never get any one to agree on the of colour of the bicycle shed
** but feel free to roll your own especially if you feel someone else's APP doesn't cut it

Will Deakin 3d ago

@stefanlindbohm @quin_antarctic @cycling_on_rails @erpu @jon @bovine3dom how do I know?

FWIW I have spent the best part of five years on the change control management board for these TAF and TAP TSI* messages and it is a key requirement that the path identifier and operational identifiers are sent in each message and that there are an additional messages that are used to link paths and train services.

2/n

* the forerunner to TEL TSI until February this year

Cycling on Rails 3d ago

@wnd @stefanlindbohm @quin_antarctic @erpu @jon @bovine3dom IMO open-data sharing is enough, as it enables 3rd parties to innovate and present it in a nice UI, ultimately overtaking the incumbent apps or pushing the incumbents to make improvements themselves.

But it works best when all the data is open (timetable, delays, train composition, ticket prices, etc.), otherwise the incumbent has a competitive advantage that shadows 3rd parties.

Oliver Blanthorn 3d ago

@cycling_on_rails i think we need more requirements for dog fooding open data because it is often crap compared to what they use internally

@wnd @stefanlindbohm @quin_antarctic @erpu @jon

@bovine3dom @cycling_on_rails @wnd @stefanlindbohm @erpu @jon Why not decide on (European) standard for how this data should be supplied?

And yes, I can imagine the "decide"-part of this statement is probably a challenge.

Will Deakin 3d ago

@quin_antarctic @bovine3dom @cycling_on_rails @stefanlindbohm @erpu @jon yes. It is called the Telematics TSI which supercedes the TAF and TAP TSIs.

The issue with the latter is that there were weasel words about sharing of open data...

Oliver Blanthorn 3d ago

@wnd December 2028 iirc?

As far as I know people (infrastructure, operators, ticket sellers) will be obliged to use the data

@cycling_on_rails @quin_antarctic @erpu @jon @stefanlindbohm

Will Deakin 3d ago

@bovine3dom @cycling_on_rails @quin_antarctic @erpu @jon @stefanlindbohm there is a shouting match* about that whether that is fair and what happens if I don't hit that date? But yes, that sounds about right.

* le plus fort dans une langue romane populaire du jour...

Oliver Blanthorn 3d ago

@wnd after three years of non-compliance the commission will issue an opinion saying that said parties should consider working towards starting implementation of the regulations

Will Deakin 3d ago

@bovine3dom @cycling_on_rails @quin_antarctic @erpu @jon @stefanlindbohm as someone who does this stuff as a job, we mostly do today. The difference is that there is now a presumption is this data is made open and publicly available.