[Talk-GB] UK Rights of Way - WikiProject
Andrew Chadwick
a.t.chadwick at gmail.com
Thu May 3 12:58:22 BST 2012
On 02/05/12 16:41, Nick Whitelegg wrote:
>>One project goal might be to consolidate the various scattered
>>information on the wiki describing how to map RoWs in the first place.
>>Come up with *one* consensus approach. We seem to be settling on
>>designation=* + highway={foot,cycle,bridle}way, by the looks of it (full
>>disclosure; it's the approach I'm cheerleading).
>
> Contentious point :-) Many, myself included, prefer the
> highway=service|track|path|unclassified plus designation=whatever.
We both agree on using designation. This is good.
Would you also agree that h=paths are generally too narrow to use in a
4-wheeled vehicle? After all, that's what h=tracks or the other road
types are intended for.
By now, h=footway seems merely a specialisation of h=path. The _only_
information it adds is that it's normally used by pedestrians, or that
it is built to be used by them. Using the more specific tag conveys
useful information information about the footpath's place in the
transportation network. The same sort of specialisation applies to
h=bridleway and h=cycleway.
Tagging it like this alone is both more useful and less vague than
h=path on its own, and we should be striving for expressiveness when
tagging. I'm not sure why you'd want to use h=path as the primary choice
really. Granted, the definition above has a disjunction, but adding
explicit tags for the physical characteristics or access tags fixes that
up nicely.
There is no such thing as a mappable path which is neither used by
anyone nor currently built up for use by someone.
FWIW, I'll still use h=path for true armchair "dunno" cases. It's by no
means a useless tag, but to me it signifies that someone, myself
included, hasn't gone and looked and made the distinction.
> This way neatly separates out the physical characteristics of the way
> and its rights,
Well, h=path has an implicit vehicle width limitation, and so do the
more specific types of it. I'm with you on representing access
explicitly; I don't think you should be inferring anything more than no
motor vehicles for all kinds of path and a single <foo>=yes for a
highway=<foo>way, and even then you want to be holding your nose as you
do it.
> and allows multi-layer rendering such as
> that done on Freemap.
Not sure what you mean by this. Could you clarify? If it's just
designations that make this happen, I think both systems work.
> I do admit to using highway=bridleway as well, but purely to "tag for
> the renderer" and make bridleways appear in a different
> colour on the main Mapnik renderer.
You're also being more concise and specific when you do this, so thank
you! And of course: tagging for the renderer (yawn!) is not a bad thing
provided you're not mapping for the foibles of a particular renderer.
--
Andrew Chadwick
More information about the Talk-GB
mailing list