[Tagging] tagging the multipolygon model (was landuse and military)

[sly (sylvain letuffe)]

On jeudi 15 octobre 2009, Pieren wrote:
> Are you suggesting that when a landuse is inside another landuse, we
> just don't use a multipolygon relation and don't care if the big is
> overlapping the small ?

Not suggesting, exploring solutions. I'm probably missing elements, but I feel 
it could be possible, and a much lower burden on mappers.

But, yes. Explore the idea of putting an end to inner polygons when something 
exist in this inner.

> But we are not modelling the reality in this case. 
why not ?
"(the matrix) It is the world that has been pulled over your eyes to blind you 
from the truth."

what if the truth should be considered a derivate from the osm database ? A 
filter ? If well define, the modeling would contain reality. (And ease 
mappers's pain)


> If someone or some 
> software requests only a certain type of landuse, it will get a full
> polygone without any hole which is not correct. 
exact

> By doing this, you 
> force all softwares to ask for all data and make complex calculations
> just to find what has to be excluded. 
exact, but not "just to find", in the main goal to ease mapping.


> I'm usually not a supporter of 
> 'tagging for the software' 
Yeah, you surprise me ;-)

But my point is also toward another consideration :
Mappers will tend to walk that way, and nothing can stop them. You can't force 
a mapper to add a thousand small lake in a forest to an inner polygon. Let's 
now ease our pain and accept it. software are allready forced to do some 
calculation (order in mapnik rendering)


PS: pieren arguing for relation while sly is arguing against, the world is 
upside down ;-) !
-- 
sly 
Sylvain Letuffe liste at letuffe.org
qui suis-je : http://slyserv.dyndns.org