[iRail] unique id-name-latlon

[Pieter Colpaert]

Hi Flo and list,

sorry for my late reaction on this one. I think this is a very important
issue and we should indeed provide a unique ID for each vehicle.

Our ID should be: readable, unique, containing the original id,
universal.

I was thinking about something like this:

CountrycodeTypeofvehicleCompanyInternalid

for instance:
BeTrainNMBSP2000


Comments appreciated,

Pieter




On Fri, 2010-08-13 at 13:51 +0200, flo wrote:
> Hi,
> 
> really interesting discussions going on here!
> 
> Personally I really like the simplicity of the iRail API in design!
> It would be really great to have this as a worldwide available
> standard that makes transit software interoperable.
> Until we get to this point (ideal future) there is always that human
> interface HTML which is also kind of an API ;-)
> As you said it seems the "right way to go" to use that interface and
> stick with names (string) for stations for now.
> That is how humans generally interact with such "HTML-APIs" and what
> we kind of emulate with our "machines".
> Certainly most transit systems use some kind of (internal) unique
> "ids" for all things identifyable like stations and stops. In some
> countries these identifiers are even written on the visible signs for
> example at bus stops. In this case they are often mapped in
> openstreetmap with the "ref" tag. In other countries like UK there
> have been large official dataset imports to openstreetmap (eg NAPTAN)
> so that many official ids are already available there. Generally the
> OSM-id of a station/stops seems a good candidate for a global unique
> independent identifier. Most of the "names" (string) are already
> available there under a share-friendly (cc-by-sa) license. Additional
> area/agency/system-specific ids would be great to collect/have in the
> long term but maybe we should start with names and make exact
> identifiers optional where available (datasetwise and apiwise) e.g.:
> ...trains.php?from=STATIONNAME&to=STATIONNAME
> ...trains.php?from_id=STATION-ID&to_id=STATION-ID
> just my2cents
> 
> Another issue that I wonder what you think about is the
> id-name-latlon-uniqueness. In many situations there is for example a
> bus_stop at a junction that has four physical latlon places on
> different sides of the road where people actually wait and the busses
> actually do stop. But all these places share the same stop-name. At
> bigger (eg central) stations there are many different  things mapped
> like train-tracks/-platforms, subway-entrances, bus-stops, etc all
> with the same name of that bigger station. For autocompletion in a
> textfield it doesn't make sense to have the same name several times.
> also the string-parameter for HTML-APIs can be unique. But for mapping
> to hafas-ids or osm-ids or for nearby-search or for displaying stops
> on a map it makes sense to separate them as soon as the distances
> become larger than eg just the opposite of the street..?
> 
> Regarding the wiki
> http://github.com/Tuinslak/iRail/blob/master/api/trains.php is it ok
> to add teleportr.org repos even if the iRail api is not implemented
> yet? it certainly will be soon ;-)
> 
> best wishes and happy scraping!
> flo
> _______________________________________________
> iRail mailing list
> iRail at list.irail.be
> http://lists.rootspirit.com/mailman/listinfo/irail