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

Ilpo Järvinen ilpo.jarvinen at helsinki.fi
Sun Aug 2 22:58:41 UTC 2015

On Sun, 2 Aug 2015, geow wrote:

> Richard Z. wrote
> > ...
> > I would leave it alone and introduce highway=footpath which would be a
> > variant 
> > of path for pedestrians, not suited or permitted for horses and vehicles
> > unless 
> > otherwise tagged and expected to be more demanding than footways.
> > ...
> @Richard - I wouldn't even dream of that ;-) Actually - do we really need 5
> or even 6 highway types for non motorized traffic?
> Wouldn't it be better to use the universal and compatible "highway=path" 
> along with specific and unmistakable attributes for physical and access
> properties. That way we could replace all highway=footway/cycleway/bridleway
> keys.
> The mess as you described it, was partly caused by mixing physical tags and
> assumed access-restrictions in these traditional keys.

Many mappers don't want to input all those types using many keys because 
of increased effort that slows down useful mapping. They could all could 
go directly into highway=* instead to make it less effort to input the 
same amount of information (1 key vs 2-4+?).

I personally would prefer that something would really be defined into 
highway=* for real paths that are not "constructed" (and that it would 
also render with default mapnik as otherwise the feedback satisfaction 
factor won't be there and it won't fly against highway=path mess that "at 
least renders"). That would probably make the issue slightly less 
convoluted eventually (and might allow easier migration between footway 
and path or even defining eventually footway == path as someone 
suggested). ...Sadly the highway=trail discussion lead to nowhere on this 
front [1]. There's informal=yes (and perhaps wheelchair=no too) but that's 
2-3 keys with no really good reason, IMHO.

However, I'm painfully aware that also all these discussions are unlikely 
lead nowhere as highway=path only supporters seem to be unwilling to allow 
such differentiation (which, according to their claims is exactly same 
class as highway=path and therefore it would be trivial to match them in 
the data user end). I also don't believe that it would be that hard to use 
correctly in practice although some likely would try to claim that such 
highway class woule be very subjective.


[1] https://lists.openstreetmap.org/pipermail/tagging/2010-October/005417.html

More information about the Tagging mailing list