Great ! Just one question : do we have to ignore accents ? In french, you write "pres, pres de" for "near". But the correct syntax is "près, près de". Or do we have to write all possibilities "pres, pres de, près, près de" ?
Pieren On Mon, Nov 24, 2008 at 7:12 PM, David Earl <[EMAIL PROTECTED]> wrote: > I need your help please... > > (Thank you to those of you who have started already before I'd even > finished preparing the page!) > > While the name finder search currently does quite a good job of handling > nuances of names in different languages, it can't currently find > categories of things in anything other than English (British English at > that). The recent mention of trying to search for "train stations near > New York" for example, doesn't work. Nor does "gares pres [de] Paris" > (or even "railway stations near Cambridge" as I only currently do one > category term, which in this case is just "stations"). > > One of the things I want to do in a revamp of the name finder ahead of > doing a complete regeneration of the index over Christmas is to offer > alternative category terms in multiple languages, so that "gares pres > Paris" and the like does work. > > I have started a page here: > > http://wiki.openstreetmap.org/wiki/Name_finder:Translations > > to collect translations. Please help by adding your own languages and > variations to it. Anything added by December 24 that is sensible and > feasible will go in - anything later I will add with the next reload in > the middle of 2009). > > I haven't listed every feature (it is hard to imaging someone looking > for "covered reservoirs near Cambridge" and "feux pres de Nancy", for > example, though if you feel like adding equivalences for categories not > listed, please do. But remember that while this is loosely based on map > features, it is about human searchability: hence several tag > combinations might be found with one search term (e.g. supermarket for > shop=supermarket and amenity=supermarket), and vice-versa (mosque from > religion=islam or moslem or muslim). > > I wonder whether the same list might serve as a possible API equivalence > table in the future, where one can set the API to work in your own > language and it translates to and from the canonical (c/f "Barangays" > discussion earlier). But at present there are things I don't do: > highways for example are only searched by name not category (and I > propose to leave it that way because of the somewhat specialized > vocabulary and somewhat arbitrary categorization; a search for just > "streets near ..." is also not terribly helpful in its own right because > of the number of hits, but will be covered separately by a forthcoming > gazetteer facility). > > I'm also reconsidering adding "in" as well as "near" (and comma) as the > separator. The problem is that there are a couple of place names in > England which have "In" in their names - Henley-In-Arden for example. (A > second problem is that since it works only by proximity, "in" will yeild > things that are wildly not "in" somewhere, but that's outweighed by what > people naturally type, I think). > > However, I think I may use some subtle capitalization to deal with these > minor cases. But extending the separator to multiple languages may cause > some clashes, and many I wouldn't be aware of. For example, are there > any French (or any language) names of the form "X pres de Y"? > > David > > _______________________________________________ > talk mailing list > talk@openstreetmap.org > http://lists.openstreetmap.org/listinfo/talk > _______________________________________________ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk