[OSM-talk] tagging long distance footpaths (and other named footpaths)

David Earl david at frankieandshadow.com
Tue Oct 23 22:47:25 BST 2007


On 23/10/2007 21:52, Peter Miller wrote:
> 
> I am in your hands guys. I see the merit of using the relationships as this
> allows all sorts of tagging options and I am happy to try it that way. 

> If so
> we would have to agree the tag values to be used within the relationship to
> specify the correct category of path (long distance European etc).


I don't think you need to define all possible values now, so long as we 
agree that the network tag identifies the network concerned, within the 
category of the wider route type (pedestrian, bicycle, bus etc), and we 
can add new ones as they arise.

For example, if we decide that "network=uk_ldp" is the tag for the UK 
long distance path network (as per the "acorn" way markers), then you 
can go ahead and define those knowing that you'll be working in your own 
space, and that someone else can do their own country;s equivalent, or 
another network of uk footpaths independently of the UK LDPs (if there 
is such a thing. There is for cycle networks in the UK - there are 
regional ones, the NCN and the half-hearted "Hovis" ones).

> Is that easier than splitting semicolons(using the ncn,rcn type tags)?

We had this discussion a while back. The problem is you have to maintain 
a whole series of parallel "columns" within each set of tags, and keep 
them all in step.

Relationships were designed to address this problem (among others).

And the Route relationship was proposed to address this specific 
application.

> Is there really a bigger decision to make about tagging routes in general?
> and when should we make it?

Like adopting the proposal you mean?

> The general motto of 'render and they will come' may apply. Possibly I will
> wait until someone starts rendering this stuff and then follow their lead.

Indeed. Rendering the semicolon variety is a nightmare. Rendering a 
relationship, e.g. by putting green blobs along the relevant ways (like 
OS does the NCN) ought to be viable. A simplistic easy approach would 
be: for each relation you would locate each of the ways and draw 
whatever is wanted on the line of the way (possibly on top of what was 
already there by virtue of the way itself - e.g. the green blobs on top 
of the road shading and casement)

David





More information about the talk mailing list