Iván Sánchez Ortega
ivansanchez at escomposlinux.org
Tue Jul 24 22:15:10 BST 2007
El Martes, 24 de Julio de 2007, Steve Coast escribió:
> So, my question is what's the status of building=* and what is the
> status of mapnik and t at h rendering it?
AFAIK, mapnik doesn't render buildings, but t at h does.
> If those two or three things line up, should we start putting in the
> outlines of lots of buildings?
We (spanish guys) have already put a number of *important* buildings:
monuments (castles), administrative (courthouses), or otherwise important
I suggest to delay *massive* building tagging (i.e. every building possible)
until we have a way to accomodate cadastrial data *and* touching, ordered
For the first point, put this into your .josm/preferences file, get any area
of urban spanish soil, zoom in a great deal, and get an image:
wmsplugin.url.7.name=Spain cadastrial data
You'll see how many *stories* (and balconies, and underground floors) every
building has. How can we deal with that kind of data?
The second problem is more grave IMHO: according to the actual standards,
there is no way to correctly *draw* three touching (adjacent) ways.
I'll explain myself: the current data model (and validator rules) encourage
- Reuse data (reuse a node if possible)
- Reuse a segment if you really, really need to
- Closed ways "areas" must be ordered. Either it draws a clockwise polygon, or
a counter-clockwise polygon.
When drawing buildings, we "really really" need to share segments between
ways: those buildings are touching, indeed.
Now, with that in mind, it's impossible to draw three adjacent areas,
maintaining "clockwiseness" in evert one.
Warning: crappy ascii-art follows. Switch to a fixed-width font:
| | |
| | |
Gentlemen, start your flamethrowers...
Iván Sánchez Ortega <ivansanchez at escomposlinux.org>
"You can get more with a kind word and a gun than you can with a kind word
- Al Capone (1899-1947)
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 189 bytes
Desc: This is a digitally signed message part.
More information about the talk