[Talk-GB] How do I tag this cyckeway?
gravitystorm at gmail.com
Thu Aug 16 15:23:38 BST 2007
On 8/16/07, Matthew Westcott <matthew at west.co.tt> wrote:
> On 16 Aug 2007, at 14:36, Rik van der Helm wrote:
> > I doubt whetter Dave's whole 'route' fits in 'highway=cycleway', where
> > it is reserved for 'cycleways exclusive for cycles'. His posting makes
> > clear the route is mainly shared use. So Dave will have a hell of a
> > job
> > defining all those highways he's running along like
> > residential/service/track/tertiary/bridleway.
> I'm not so sure that you can always piggyback cycle routes on top of
> existing 'highway' ways anyhow - I've encountered NCN routes and
> signposted walking routes that follow a residential street for some
> its length and then turn off down a side alley. This makes me think
> that ultimately the most appropriate solution might be to define a
> new way for each designated route, sharing the same segments as the
> underlying roads - or at least it would be if our editors were better
> at representing / editing overlapping ways...
Currently, I would mark your scenario by having two residential
streets and one side alley - the residential streets would have the
same name, but the appropriate one would have an additional ncn_ref
If entities were in use, there would be the same layout of ways.
However, you could have a 'Example street' entity made up of the two
residential ways to show that they were linked (and only put the name
on the entity), and another entity to represent the cycle route, which
would have one residential way and one side alley as its members.
Cycle *routes* (like the ncns) are just a logical combination of
sections of roads, footpaths and so on that make up an overall route,
so they will always consist of a collection of underlying highway
types. *Cycleways* are physical stretches highway designed especially
for cycling on.
I don't see anything wrong with 'piggybacking' logical routes on top
of physical highways. If we did them as completely separate ways
(sharing common segments), it gives us few advantages over the current
methods (and I can't see any advantage over using entities) and makes
life much harder.
More information about the Talk-GB