[OSM-talk] Collapsed names
david at frankieandshadow.com
Sun Jul 27 12:14:42 BST 2008
On 27/07/2008 11:00, Igor Brejc wrote:
> David Earl wrote:
>> In principle it ought to be possible to determine the country an
>> object is in, even though it is quite hard at present. I think a lot
>> of things would benefit from this ability: nationally-styled rendering
>> rules, deciding which way roundabouts go, name renderings, validation,
>> improved searching context etc.
> I think it is not as easy as it looks. Some countries (Slovenia, for
> example) still do not have the whole border drawn in OSM (lack of free
> data), so you cannot use this to determine what belongs to what country
> - never mind the purely technical difficulties of doing so.
Well, depends on the application. And like all our applications, they
fail with incomplete data, but that will be rectified over time. Even an
approximate boundary will catch most of the area of a country.
> And anyway,
> even this does not help for multilingual countries like Belgium or Spain.
Well, depends on the application - choosing a national rendering would
be OK. Bilingual countries should be using name:lang widely as you say
below, and it may be that the country it is in could define the default
language for the unadorned name= tag (so are places in Catalunya tagged
in Spanish by default, with name:ca for Catalan; if not, how does one
know what language name= refers to).
> It is probably more logical to use multilingual place names which should
> have an ISO language ID, like "name:ca" so that we can match them with
> appropriate language mappings like Name finder:Abbreviations. But this
> multilingual tagging is not used universally in the OSM community.
name:lang does seem to be quite widely used. What you're suggesting
would be quite a good idea except that I don't have any way of knowing
what the language is of the simple "name=" - I can deal with the
alternates, but not the principal name. And with millions of names now
in there, these ain't going to get manually qualified (and that would
have to be by a new tag or duplication because of the way name is used
now for rendering), so unfortunately it is not likely to be practical.
> There is also an "is_in" tag, but again, it's not widely used (at least
> to my knowledge, I could be wrong).
It's widely used in the UK. It would be better IMO if it was categorised
(is_in:country=Engl;and, is_in:county=Cambridgeshire for example) or had
some hierarchical structure (but that's difficult to maintain), but as
it is so ingrained, again it isn't likely to change any time soon.
> BTW: it would be good to add ISO language IDs to the Name
> finder:Abbreviations page so that the mappings can be computer-matched
> more easily.
Feel free. I'm not parsing this list automatically anywhere at present,
I'm just using it as a manual reference for a table in the name finder.
More information about the talk