For the first problem I have often used an approximate approach, in the
sense that if there were two different-level "give-ways" I would put the
more severe one, i.e. the stop.
In other cases I have created two ways taking the painted divider as a real
one.
For the second problem, I have simply ignored it.

I agree all three are not satisfactory.

The relation solution seems to be correct approach to me, even though it is
little-used.
There are 59 stop or give-way relations in the US, but 57 of them are in
the same city.
But there are >1392 relations with restriction:bicycle=give_way in France
and the Relation restriction Wiki page
<https://wiki.openstreetmap.org/wiki/Relation:restriction>
lists it. They ave popped into existence with the turn-right-on-red for
bikes regulation, which is spreading in Europe.
So it would be a small extrapolation from there to having generic stop and
give-way restriction relations.


On Sat, 2 May 2020 at 22:29, Jarek Piórkowski <ja...@piorkowski.ca> wrote:

> On Sat, 2 May 2020 at 16:21, António Madeira <antoniomade...@gmx.com>
> wrote:
> > I'm not very knowledgable about relations, and I'm sorry if I'm a bit
> confused here, but doesn't a restriction relation means the exact opposite
> of what's intended here?
> > I mean, I want to apply a STOP sign to a given lane (in a way with two
> lanes, for example) and force its action to a given direction on the new
> road ahead.
>
> IMO, a relation helps here because you can define the route which the
> rule applies to - it only applies going "from" a certain way and "to"
> a certain way, and specifically applies at a given "position".
>
> Then it is just a matter of choosing what type of relation works best,
> or creating a new type.
>
> > If neither relation scheme (enforcement or restriction) can be applied
> here (for complexity or incompatibility reasons), why not use the existing
> lanes scheme?
> > Like this:
> >
> > highway=stop
> > stop:lanes=yes|no
> > stop:turn:lanes=left
>
> Personally I've always seen highway=stop on a node, and I'm not sure
> :lanes tagging makes sense on a node (a point doesn't have lanes).
> You'd definitely need to add at least direction=forward/backward if
> tagging a node. But I wouldn't be opposed to that scheme in general.
>
> --Jarek
>
> _______________________________________________
> Tagging mailing list
> Tagging@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/tagging
>
_______________________________________________
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Reply via email to