[OSM-talk] colour fills in osmarender

[Etienne]

On 6/9/06, Richard Fairhurst <richard at systemed.net> wrote:
>
> Andy wrote:
>
> > I don't like the fact that I create tree-like ways either. The problem
> is
> > that the trees often have short splitting branches at the ends
> representing
> > the little turnouts you find in the real world. Creating one of these
> extra
> > bits as a separate way for each branch that splits is cumbersome.
>
> I agree, but that's a UI problem, rather than a data model issue.
> Editor software should ideally have an 'reapply last set of
> keys/values' control, so you could instantly create another way with
> the same set of attributes as the last one.


It is not a UI problem.  It is a data modelling problem.  If several ways
are all part of the same street then the data model needs to maintain some
link between these items otherwise you end up with duplicated data and
various other problems.


> Another problem I come up against is the run into and out of roundabouts.
> > Nowadays many of our bigger roundabouts have a divided section of
> normally
> > single carriageway road (separated by an island) on the run in to a
> > roundabout.
>
> Hm, yes. I was wondering about these when doing the Worcester ring
> road. My personal conclusion was that if we don't do them for islands
> in the middle of the road at (say) a pedestrian crossing, we shouldn't
> do them for roundabouts, either.


Aparrently these things are called Flared Entries or Flared Approaches (
http://www.standardsforhighways.co.uk/dmrb/vol6/section2/td1693a.pdf -
Section 2 is very interesting - should be required reading for OSMers).

Where I know about them I have always drawn them in.

> So I guess my thinking leads me to two separate options. Either I need a
> > separate receptacle that carries a collection of elements with the same
> > attributes which is not a "way" (or I use a way anyway as I do now) or
> for
> > anything to do with routing and path analysis I ignore any non-linear
> ways
> > and use the underlying node and segment data instead?
>
> Option 3. You decide to do everything through individual ways, and
> smile sweetly at Imi (and other editor authors) asking him to include
> a "reapply same tags" feature.
>
> cheers
> Richard
>
>
> _______________________________________________
> talk mailing list
> talk at openstreetmap.org
> http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk/attachments/20060609/7604099c/attachment.html>