[OSM-talk] RFC: render explicitly used "oneway=no"

Stanislav Brabec utx at penguin.cz
Mon Sep 1 19:36:34 BST 2008


Tordanik wrote:

> > Alternate proposal:
> > New tag "render_oneway=yes" or new value for "oneway=no:visible".
> 
> So it's plainly a rendering hint for a certain set of applications,
> right?

Yes. Currently there is no way to specify, that the highway layout is
very complex and the standard "one arrow per defined distance" may be
confusing. See the link in the first mail for an example - one street
has just only four segments, but each segment has a different oneway
status (up, bi-dir, down, up). In Prague downtown there is a lot of
streets that change oneway status in the middle or even reverse the
direction in the middle (and one has to turn there).

> (Not for all of them, e.g. not for a renderer that uses colors to
> mark a oneway street. A renderer could also get most of these cases
> algorithmically.

Actually, very hardly (find all objects connected to the same node,
which either have the same name or they are continuing in an angle near
to 180 degrees. If these objects have oneway tag and the part in
question does not, then the part in question should be rendered with
bi-dir arrow.).

In future it may be simpler with Segmented tag:
http://wiki.openstreetmap.org/index.php/Relations/Proposed/Segmented_Tag

> ) I do not really like this sort of tags, but I'd at
> least suggest that you explicitly state that it is application specific
> by using a "osmarender:" or "rendering:" prefix or something similar for
> the key.

OK. What aboutthe next proposal:

"rendering:oneway=yes"

Rendering hint: The way has complex layout. "oneway" tag needs precise
rendering. For example: Render arrow in each path segment, render double
arrow for segments with "oneway=no".


________________________________________________________________________
Stanislav Brabec
http://www.penguin.cz/~utx





More information about the talk mailing list