On Tue, 03 Mar 2009 00:15:30 +0100, Frederik Ramm <[email protected]>  
wrote:
> Hermann Kraus wrote:
>> I already noted this in a comment:
>> # In theory each way could be represented as a multipolygon with the
>> # following properties:
>
> I saw this comment; I believe you wanted to say "each *closed* way"?

There is nothing that enforces that the inner or outer ways are closed. I  
only display a warning message when my code can't form a closed way, but  
that could be removed.

My multipolygon object is very similiar to the current way object, with 2  
major differences:
- There is no 'nodes' element.
- There is a 'outer' and a 'inner' element which contain a list of ways.  
For the 'outer' element there is no need that it is a list of ways, it  
could a well be a list of connected nodes ([[nd1, nd2, nd3], [nd4, nd5,  
nd6, ...], ...]) because all the tags are stored in the multipolygon  
object. The 'inner' list would be replaced by a list of "multipolygons"  
(the complete object would have to get a new name of course, because it no  
longer has to be a multipolygon at all).

But as I already wrote this would require a lot of changes. I don't think  
it is worth the effort just to have a little bit cleaner code. When we get  
a native area type as Frederik mentioned everything will have to be  
changed again anyways. And I _really_ hope we get a native type soon!

Hermann


_______________________________________________
Tilesathome mailing list
[email protected]
http://lists.openstreetmap.org/listinfo/tilesathome

Reply via email to