[iRail] Help needed: stationslist & wikipedia

[Pieter Colpaert]

Hi,

To summarize the problem that we're having:

That list you showed us is what we currently have but the matching
between railtime ID's, geocoordinates, NMBS ID's (these differ from
railtime id's) and English names is really hard and currently a little
buggy. 

Since the API also allows people to loosely match a station, such as
give in "ghent" for Gent Sint Pieters, sometimes weird results are
returned and that's the problem we're trying to fix.

Christophe also wants a better stationslist returned at
api.irail.be/stations/. Without parameter this list should not return a
list of all stations with their English names (this list is the most
buggy of all since there is no official "english" name), but it should,
according to Christophe, return a list of all *stationnames* (not unique
stations) which are different.

For instance, if we have Charleroi ouest, returning charleroi west as
well would be crazy since everyone can understand the stationname and it
would be overhead to add it to an autocompletion list since it will only
make it dirtier. If however we have Doornik, we also want to return
Tournai, since these are 2 totally different names. Even for Oudenaarde
this is an issue, since French users will start typing Aude...

To be frank, I have broken my head over this the last year and the
implementation we currently have is the best I could do. I think it's
cool that now we are arguing about this because we may find a better
solution.

Your sincerely,  

Pieter

On Mon, 2011-08-15 at 10:36 +0200, Tom Waterbley wrote:
> Why don't you try to parse this: 
> http://www.railtime.be/website/StationDataScript.ashx
> 
> --------------------------------------------------
> From: "Pieter Colpaert" <pieter.colpaert at gmail.com>
> Sent: Monday, August 15, 2011 2:06 AM
> To: "iRail" <iRail at list.iRail.be>
> Subject: [iRail] Help needed: stationslist & wikipedia
> 
> > Dear List,
> >
> > As Christophe brought to our attention earlier today we need a better
> > stationslist @ http://api.iRail.be/stations/ . It really would be a
> > shame if we would do double work or if we would keep all our efforts
> > inside the NPO. I started thinking and now I came up with this solution:
> >
> > Wikipedia is pretty awesome. We got all (really, ALL) information that
> > one could imagine about any (really, ANY) station on it! Check this out:
> >
> > http://fr.wikipedia.org/wiki/Cat%C3%A9gorie:Gare_de_Belgique
> > or
> > http://nl.wikipedia.org/wiki/Categorie:Spoorwegstation_in_Belgi%C3%AB
> >
> > The English version however is not sufficient at all
> >
> > http://en.wikipedia.org/wiki/Category:Railway_stations_in_Belgium
> >
> >
> > DBPedia
> > -------
> > DBPedia is a system where you can query wikipedia as if it was a
> > relational database. This means that we can just generate a stationslist
> > by asking the right thing. I made a query:
> >
> > select *
> > where {
> > ?station a yago:RailwayStationsInBelgium
> > OPTIONAL { ?station dbpprop:code ?code }
> > OPTIONAL { ?station geo:lat ?lat }
> > OPTIONAL { ?station geo:long ?long }
> > }
> >
> > And put it into http://dbpedia.org/sparql
> > Which gave this result: http://dbpedia.org/sparql?default-graph-uri=http
> > %3A%2F%2Fdbpedia.org&query=select+*%0D%0Awhere+{%0D%0A%3Fstation+a+yago%
> > 3ARailwayStationsInBelgium%0D%0AOPTIONAL+{+%3Fstation+dbpprop%3Acode+%
> > 3Fcode+}%0D%0AOPTIONAL+{+%3Fstation+geo%3Alat+%3Flat+}%0D%0AOPTIONAL+{+%
> > 3Fstation+geo%3Along+%3Flong+}%0D%0A}&debug=on&timeout=&format=text%
> > 2Fhtml&save=display&fname=
> >
> > This would give us a HUGE amount of flexibility. By adding different
> > parameters such as "?station dbpprop:abstract ?abstract"  you can even
> > get the history of a station.
> >
> >
> > PROBLEM:
> > ---------
> >
> > The data which dbpedia returns can only be retrieved from the English
> > wikipedia. This means we will not have the great extended pages like on
> > the Dutch or French version.
> >
> > However, if we, the iRail community, would just translate the Dutch or
> > French pages into English, we solved our problem: we'd have a very good
> > stationlist in our API. On top of that, we will also have made a lot of
> > other people happy who would be trying to look up a wikipedia in English
> > about a certain station.
> >
> > MISSION:
> > --------
> >
> > Let's make this an iRail issue! Let's translate all pages into English!
> > This would be a great achievement for our community.
> >
> > What do you think? Who's in?
> >
> > Yours sincerely,
> >
> > Pieter
> >
> >
> > -- 
> > iRail vzw/asbl
> > +32 (0) 486/747122
> >
> > _______________________________________________
> > iRail mailing list
> > iRail at list.irail.be
> > http://lists.rootspirit.com/mailman/listinfo/irail
> > 
> _______________________________________________
> iRail mailing list
> iRail at list.irail.be
> http://lists.rootspirit.com/mailman/listinfo/irail

-- 
iRail vzw/asbl
+32 (0) 486/747122