[OSM-talk] Feature Query update

[Christopher Schmidt]

On Sat, Dec 02, 2006 at 03:52:18PM -0000, David Earl wrote:
> This is such an enormous improvement over any rendering I've seen before,
> notably in its handling of text, how recent it is and how detailed the tiles
> are, and degredation with scale. I'm very excited to be able to see my work
> presented in such a nice way. This is the first time I've really been able
> to look at a quality detailed rendering.
> 
> I notice that areas for natural=water and amenity=parking are rendered as
> Ways, however (reservoir seems ok and as you say parks are also ok). A chunk
> of railway line in my area is also not rendered in the detailed tiles,
> though it is there in the smaller scale ones.
> 
> Both these artefacts are visible at
> http://labs.metacarta.com/osm/?lat=6802000&lon=18000&zoom=14&layers=B
> 

Anything which you consider a rendering bug, please file under the
instructions in:

  http://lists.openstreetmap.org/pipermail/talk/2006-November/009044.html

I'm about to get started going through these bugs in a couple hours,
and if yours is in there, I'll include it in the things I consider
taking care of.

> I think there could be greater differentiation of areas of different kinds
> by colour, but that's a minor quibble.

Minor quibbles that everyone has are valid quibbles :)

> Perhaps if you haven't implemented a particular area rendering, you (not
> sure who I am addresing here: is this the mapnik part) could still render
> generic types (e.g. amenity) which could be areas as opposed to ways it as a
> grey filled area rather than a way.

The osm2postgis translator only converts the types of ways it knows are
areas to be 'polygons' which show up in the rendering. Teaching it to be
more intelligent about this is one of the things on my todo list this
weekend. However, it is likely that until the tool knows about new types
of areas, they simply won't show up. I'll have to see how the code
looks. 

> IE7 also has a Javacript error and I can't get the search box to in IE7
> (this is probably that error).

I don't have IE handy to test anymore. I might be getting a copy later
this weekend, but I might not. If I do get a copy before Monday, I'll
try to test it out -- I don't see anything off the top of my head that
would cause a problem. If someone in IE can get a bit more information
such that I can debug with it, I'll gladly take a look. (I'd like to
have it work there -- not having it work in IE is embarrasing, but
unfortunately I don't have the resources to test with.) 

Thanks for the feedback.

Regards,
-- 
Christopher Schmidt
MetaCarta