[OSM-talk] the ref:color schema

Iván Sánchez Ortega ivan at sanchezortega.es
Sun Apr 19 16:20:09 BST 2009


El Domingo, 19 de Abril de 2009, Greg Troxel escribió:
> I am also starting to think that the US proposal to do "us_i_ca" for
> california interstate is maybe not the right answer.  The road really is
> "network=us_i" but maybe needs a "network:sign_variant=us_i_ca" tag.

I think that usage of "network=*" is the right way to go (indeed, in Spain, 
every road *is* legally part of a network with a name). But keeping track of 
*all* road networks and their signage is going to be difficult.

> As an example, near me in the US you can be on a road which is both I-95
> S and Massachusetts state route 128 south.  Then there is an
> intersection where I-95 splits off.  After the interchange you are on
> 128S and I-93N.

Hm, sounds a little bit like the "Nudo de la Paz"[1] here in Madrid. You can 
be either in the M-30, M-11, A-1 or M-607 depending on the *lane*.

[1] http://www.openstreetmap.org/?lat=40.48261&lon=-3.68427&zoom=15

But, yeah, relations are IMHO the way to go.

> > - Sometimes, there are more real administrative levels than OSM
> > administrative levels (i.e. yellow vs. purple tertiaries in Soria).
>
> That sounds like something to be fixed, although it seems
> primary/secondary isn't fully about administrative levels

OK, but how do we fix this?
- Do not use administrative levels for highway=* at all, as this info is 
somewhere else (network=* or even ref:color=* plus country). But then, what 
is highway=primary for??
- Use network=*, then assign highway=* depending on the importance of the 
network (or motorway for physical motorways), then get the shields through a 
batch process and a huge table.

> > - That table can be a *real* mess.
>
> Sure, that's the best argument for needing to encode colors.  It just
> seems like if one can encode the administrative level etc. and map to
> colors that's far better - maybe there needs to be a databsae of the
> mapping that all renderers and other users can get at, sort of logically
> part of the database but not a node/way/relation.

Do you really think we can work that network->color table out, and make 
software that fits into osm2pgsql to derive the shields**?


** I don't like the word "shield". We use frakking rectangles.


Cheers,
-- 
----------------------------------
Iván Sánchez Ortega <ivan at sanchezortega.es>

MSN:i_eat_s_p_a_m_for_breakfast at hotmail.com
Jabber:ivansanchez at jabber.org ; ivansanchez at kdetalk.net
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 197 bytes
Desc: This is a digitally signed message part.
URL: <http://lists.openstreetmap.org/pipermail/talk/attachments/20090419/fa54203e/attachment.pgp>


More information about the talk mailing list