[Tagging] highway=footway - Advanced definition: Distinction footway vs path

Ilpo Järvinen ilpo.jarvinen at helsinki.fi
Fri Aug 28 23:46:19 UTC 2015

On Fri, 28 Aug 2015, geow wrote:

> Ilpo Järvinen wrote
> > On Fri, 28 Aug 2015, Warin wrote:
> > 
> > For those who would want to have a separate tag for 'trails', it's 
> > exclusive but obviously those who would want to tag everything with 
> > highway=path+subtags will disagree :-).
> > 
> > I see there are three main sides in this, those who would want...
> > 
> >   1. ...highway=path == highway=footway
> >   2. ...highway=path only
> >   3. ...to have additional highway=trail (or similar value) that is 
> >      distinct from path/footway
> An additional highway type "trail" is no solution to the problem of missing
> decent classification criteria "path vs footway"

While I don't know what you exactly refer here with "path" but I assume 
you meant path not highway=path :-), I'd personally define it as a 
way that surface is generated through wear/erosion as people walk there 
but exclude those where a constructed way surface is "damaged" by 
wear/erosion. In addition, some constructions where the purpose is for 
example erosion/hazard prevention would still fall under "path" although 
the surface would not be erosion generated (e.g., duckboards). Also, some 
hiking trails that exists but are not observable on ground would need to 
be included but that's unrelated to the classification criteria between 
"path vs footway".

I think this would match very well to what people conceive as path (I 
don't refer to highway=path here) and is based on something that is 
possible to determine on ground. Obviously there are some borderline cases 
where it's close to impossible to tell if the "path" was really made or 
not (e.g., some fine_gravel added or not to a short-cut, could be carried 
by shoes too), however, for those it usually matters little whichever of 
the alternatives (=trail or =path/footway) gets used. Then there are some 
cases (although I don't think that many) where a constructed way is left
to detoriate until environment takes over entirely, and at some point it 
would have to turn into =trail but it will end up somewhat subjective (I 
find this the only case that is seem really tricky and I can understand 
you might say it's not "decent" enough because of this corner case :-)).

> and would likely cause further trouble for many sorts of reasons. 

But not for you since you don't care (as just put them all into 
highway=path instead like you admit below)? Why not let those that
care deal with that "further trouble" that shouldn't matter to you
in the first place?

> I personally prefer to use the universal and compatible "highway=path".
> Along with specific sub-tags for physical and access properties it's the
> Swiss army knife for non motorized traffic.

Is your opposing only about that you'd want to force other people to map 
what matters to you (i.e., those physical and access tags) by preventing 
them to define something that suits them (even though that doesn't 
directly harm your tags nor contradict with them)?

And btw, I don't think that =trail itself undoes anything in the use of 
the physical tags. However, I'm yet to see a rendered map which would 
cleverly visualize even somewhat complete subset of those "sub-tags" 
which people are promoting. Of course such visualization won't matter to 
all, especially to those who are from today's 
I-use-mobile-device-for-navigation people who depend on CPU processing all 
the routing from them, but there are plenty of people who still want to 
see a visualized map and route in their head. Perhaps some complex set with 
ANDing from those subtags could mimic what a simple trail/path vs footway 
distinction offers to such people but I'm yet to see that ;-). I know you 
might not agree with me but it won't change my personal experience. It 
has worked surprisingly well when route planning unknown places abroad 
(but I've ended up only visiting places where the distinction has been 
quite well maintained so there might be places where OSM quality is 
inferior to the non-OSM maps that often lack sufficient accuracy to 
allow useful route planning). There's one exception though, those funny 
highway=path+bicycle=designated downhill MTB trails in Vancouver North 
Shore Mountains that rendered like ordinary cycleway but with 
highway=trail I could have differentiated even those trivially ;-).


More information about the Tagging mailing list