[OSM-talk] Advanced relationships
frederik at remote.org
Thu Jun 14 07:27:03 BST 2007
> A turning restriction is just a link between two segments with a
> restriction applied - a segment pair + restrictions.
GDF models it as a triple (segment, node, segment), just in case you
have segments meeting at more than one node ;-)
> The only remaining question is how to represent the segment pair. My
> suggestion is to use a way, since that is the only current way to link two
> segments. A new way tag would be needed.
Come on, it isn't as if we need to express everything there is in ways
or segments or nodes!
> I know there are probably better and cleaner methods for implementing
> routing, but this method would give a fully expressive method using only
> two new tags and no change to the OSM data model.
Yes. There are numerous other extremely ugly methods how we could
utterly wreck our data without changing the data model one bit (e.g.
adding a tag to a segment "no_right_turn_into=<id of other segment>").
In my eyes this cannot be the way forward. The current data model is not
god-given. It served well for a start, but we cannot, should not, and
will not stay with it for eternity. It just isn't sufficient, and trying
to express everything in ways and segments, while seeming a smooth
transition at first, will only deform the whole thing into something
> It would not require any
> immediate change to renderers, or to editors.
Oh yes it would, because your little "relationship ways" would be very
difficult to handle in the existing editors, and any but the few
initiated folks will be tempted to just delete them.
> It is certainly considerably
> cleaner than additing additional oneway ways that don't really exist.
I hesitate to use the word "clean" in conjunction with such a hack.
> I know the routing issue comes up frequently, and the response often seems
> to be that it is too hard for now. I think it needs focus, as I think it
> is significant to the success of the whole project, otherwise we risk
> ending up with a map of the world that is just a pretty picture.
Yes but little patches and hacks are not the way forward. This will end
in a data set where half of the ways and tags are not really ways or
tags but just some kind of helper constructs to circumvent the weakness
of the data model.
We need proper relationships and not a clever abuse of existing mechanisms.
Frederik Ramm ## eMail frederik at remote.org ## N49°00.09' E008°23.33'
More information about the talk