Hi,
if only I knew that the idea would arouse so intense emotions! Ok, now I
know that I should be more careful with the word "semantic" next time.
Andy Allan píše v St 20. 05. 2009 v 14:48 +0100:
> On Wed, May 20, 2009 at 2:26 PM, <marcus.wolsc...@googlemail.com> wrote:
> > On Wed, 20 May 2009 14:04:36 +0100, Radomir Cernoch
> > <radomir.cern...@gmail.com> wrote:
> >> MP píše v St 20. 05. 2009 v 14:16 +0200:
> >>> I wonder, can we have at some place (wiki?) some definition file that
> >>> will specify these per-country default limits in some machine-readable
> >>> way?
> >>
> >> Yes, surely! My personal idea of "best solution" is to use the Semantic
> >> wiki, because it provides both machine-readable and human-readable
> >> format in one place.
> >
> >
> > Agreed. This sounds like a very good idea.
>
> I think it sounds like a terrible idea.
>
> A) We have a geo-database for geographic information
> B) We have a wiki for project-support information
>
> Why would this particular geo-data not live in the Geodatabase?
>
> Let's take the very first bit of the example:
>
> country(cz) {
>
> How can I tell if a particular way is in the country "cz"? Maybe I
> need a lookup table that gives me a list of coordinates to specificy
> the boundary of that country. Should this be in the wiki too? Do we
> have somewhere better for storing lists of coordinates?
Of course not.
> So let's make the assumption we have an object (I'm guessing a
> relation would be handy) for each country explaining where it is. Now
> if only there was a way of assigning attributes to the country to hold
> information that applied to that country. Geez, maybe we could allow
> tags on relations?
I do not mind putting all this into the relations for each country.
However the question is, how complicated the whole system will be.
If we end up with 'speed_city', 'speed_rural' and 'speed_highway', then
why not. However there is a danger of ending up with tags like:
'speed_limit_on_highway_inside_polygon_tagged_place' or
'speed_limit_on_primary_not_in_any_polygon_tagged_place_or_built-up'.
This is an old problem of relational-database expressiveness. For now I
propose to wait until we collect all nitty-gritties of all different
traffic regulations we know. The discussion about implementation can
come second.
> Hmmm. Or maybe we should ditch the whole geo-db idea and just put
> everything into a wonderful Semantic Wiki.
When someone invents indexes into ontology, which would be as fast as
indexes into geo-db, why not? Personally I am rather sceptical about it.
;-)
> Cheers,
> Andy
_______________________________________________
talk mailing list
talk@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk
