search

VDVde / OJP

Open API for distributed journey planning. CEN/TS 17118:2017.
https://www.vdv.de/open-journey-planner.aspx
22 stars 12 forks source link

Empty PublishedLineName #71

Closed ue71603 closed 4 years ago

ue71603 commented 4 years ago

Hi everybody In some cases there exists no LineName. However, the field is mandatory and can't be empty. Sending a blank does not seem a good idea. Should we change the type? What are other opinions? I think we need to coordinate this also with the SIRI CEN group. I ask there, too. Kind regards Matthias

`

Tyoe for a string in a specified language. Mir scheint also, als könnte man A restriction of W3C XML Schema's string that requires at least one character of text. `
ue71603 commented 4 years ago

Or better, PublishedLineName is optional. That's the case in Siri. Then it is clear. Can we make it optional also in OJP?

pvgrumbkow commented 4 years ago

The comment of PublishedLineName states: "Line name or service description as known to the public, f.e. "512", "S8" or "Circle Line". In my point of view, this is always there. Every service has a line name or a service name, which is communicated to travelers. Therefore a traveler information API like OJP needs to have this kind of information as mandatory. Could you provide examples, where a service does not have such a communicated name?

ue71603 commented 4 years ago

In Switzerland some regional trains don't have a line. Even the lines of IC1 and S1 are not really PublishedLineNames. The correct "published line" for both (and all six S1 that exist in Switzerland) is "1". The S1 in Lausanne, Bern, Basel, Zürich and the IC1 all share a pseudo line number of 1.

The RE Schaffhausen-Chur has no line. It only has a train number: Regional-Express 3472

This is by the way true for most train operators in Europe. They use train numbers. Lines is mainly a concept of local transport. I know that we will get to lines also for trains, but currently to fill in the PublishedLineName is in general just not correct.

ue71603 commented 4 years ago

In the Swiss data and in the current restrictions, we will just but a "space" there. And I don't guess that's helpful.

pvgrumbkow commented 4 years ago

The comment for PublishedLineName states clearly, that it also may contain a "service description". For a German ICE I would expect it to be filled with "ICE 884", which contains the train number. Perhaps "PublishedLineName" is the wrong term, and "PublishedServiceName" would be more appropriate. However this element is called, it is intended to be filled with a name for the given service which is used in traveler communication. This can be "S1", or "ICE 884", or "511", or "Ringlinie", or "IC 1024". It is not an identifier. Therefore I still think, that for every service there is a name for it, which is used in traveler communication. This shall be filled here.

ue71603 commented 4 years ago

I will have a look into that. In Switzerland neither 3472 nor RE is used in communication. It is the RE from Schaffhause to Chur. And i guess that would be too much for the field.

ue71603 commented 4 years ago

After discussion: RE alone would/could be used in that case. However, in the general picture: In Siri it is optional and in NeTEx all possible equivalent elements in Line are optional as well and even the Ref to al Line in VehicleJourney is optional, so it should be optional here as well. This is already a given for consistency alone.

pvgrumbkow commented 4 years ago

Actually I don't see the necessity to follow Siri or NeTEx here. OJP is a service interface in order to give information to the traveler. Siri and NeTEx are data distribution interfaces/formats. Please have a look at the TimeLegStructure and its sub structures. If PublishedLineName was optional, there would be no mandatory element anymore, which gives the service of a TimedLeg a name. An information system accessible by OJP should be able to give a service an appropriate name to be given to the traveler. Otherwise we could end up with services without any description.