[OSM-dev] proposal to kill areas

Etienne 80n80n at gmail.com
Sat Jul 22 22:53:06 BST 2006

On 7/22/06, Immanuel Scholz <immanuel.scholz at gmx.de> wrote:
> Hi,
> > All
> > The treatment of filled areas in SVG is very simple and is described
> > in about 3 paragraphs here:
> > http://www.w3.org/TR/SVG/painting.html#FillProperties.
> >
> > I've used it successfully to rendered areas in Osmarender.
> >
> > What are the problems that need to be addressed that are not covered
> > by this?
> The problem is, that we have two suggestions to choose from: either
> specify areas as "current way structure with a user defined tag" or as
> an own data element, which is proposed to be a list of nodes (polyline).
> Your website describes how the area is defined, if you have a list of
> nodes (polyline).
> To use your suggested SVG specification for defining what should be
> drawn, the former way (consisting of a list of lines) has to be
> transformed in the latter (polyline). This is possible, but there are
> many possibilities, where an way is ill-formed.

No.  The SVG specification does not have the notion of an ill-formed area.
It simply has a rule which says whether any point is inside or outside.  If
the topology of the way is not "area like" then the resulting SVG filled
area will not be area-like.

This is like saying that if you create a rectangular area of height 10 and
width 0, then you will not get an area, you will get a line.  if you create
a way that is not closed or does not enclose a single region then you will
not get an area, but you will get something.

So, there are many possibilities that will not produce an area, but this
does not mean they are ill-formed because there is no well-formed rule.  We
can, of course, add some additional OSM well-formed rules about what makes a
way (or area) be a valid area.  But what would those rules be?  And why
would they be necessary?


So if Steve chooses to implement the "easy" and "simple" way and
> removing areas from the database, clients have to check whether you can
> draw areas out of the way.
> Additional to that, I think that the checking code will be implemented
> in the server, namely as soon as the server has to provide area-specific
> requests.
> Ciao, Imi.
> PS: I get the feeling, that this discussion is senseless as Steve, who
> is actually doing the stuff seems to not participating anymore?
> Steve? ;-)
> _______________________________________________
> dev mailing list
> dev at openstreetmap.org
> http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/dev
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/dev/attachments/20060722/fe4b2d3c/attachment.html>

More information about the dev mailing list