[OSM-talk] correctly mapping avenues
Bernd Raichle
bernd at dante.de
Mon Feb 11 15:20:02 GMT 2008
Hi,
on Sunday, 10 February 2008 08:34:31 -0800,
Karl Newman <siliconfiend at gmail.com> writes:
> On Feb 10, 2008 4:21 AM, Frederik Ramm <frederik at remote.org> wrote:
> > > Since trees lining a way/street are such a common occurence, why
> > > not have a simple additional tag to the main road.
> > >
> > > lined_by_trees=yes/no/left/right
> >
> > I'm a bit unhappy about needlessly inflating the importance of the
> > direction of ways. Long-term, I would actually like to get rid of the
> > direction and express everything in relations.
This means, that you find it necessary to have something like a
"direction" or a "side", both of this features related to a way?
But you don't want to express a direction or a side by the _implicit
order_ of the way nodes.
> > The reasons for this
> > are
> >
> > (a) the direction is too easily changed, sometimes by mistake
... because none of the current OSM editors show direction- or
side-related tags explicitly.
> > (b) there might be multiple conflicting things that rely on the
> > direction, e.g. a road that is oneway from A to B but has a
> > slope from B to A
> >
> > Anything with "left/right" in it also relies on direction. I'd prefer
> > "east/west/north/south", or using an explicit relation that says
> > "trees on the right between nodes A and B along road C".
I am against east/west/north/south because there are a lot of
ways/areas/things which do not go straight ahead.
> Okay, this thread is at risk of spinning wildly off-topic, but I've been
> thinking about this situation recently. It seems to clamor for the use of
> specialized relations that are "direction-aware". That way, if a way is a
> member of a relation and has directional properties (left/right), then the
> editors could look for those relations when the way is reversed and either
> fix them automatically or at the minimum raise a warning dialog.
>
> I also had some other ideas about enforcing referential integrity for
> another type of specialized relation (if one or more node relation members
> is required to be part of a way relation member, then enforce that rule).
> That rule could actually be enforced by the API.
>
> These specialized relations would just give some structure to the wide-open
> relation type, without implying anything about the nature of the relation.
> It could possibly be accomplished through special tags on the existing
> relation structure.
Do you have any propositions how this will look like or how this
should be done?
A few days ago I have started a new proposal for a "Segmented Tag",
which relates a set of tags to a directed or undirected part of a way
(I have called this part "segment" inspired by GDF's "Segmented
Attributes"). I have not found the time yet to finalize the proposal
adding some examples, nonetheless it can already be found in the OSM
Wiki (Relations/Proposed/Segmented Tags).
Best wishes,
-bernd
More information about the talk
mailing list