[OSM-talk] Tag:highway=cycleway inconsistency

Alex Mauer hawke at hawkesnest.net
Thu Jul 3 17:06:19 BST 2008

William Waites wrote:
> Possibly it is better to remove implication relationships amongst tags. 

I doubt it.  I think it is safe for highway=cycleway to imply
bicycle=yes and motorcar=no, and for highway=motorway to imply foot=no,
horse=no and bicycle=no.  These are "obviously correct" assumptions, as
they are part of the definition of the cycleway, or the motorway.  So I
think some implications are quite important, as they indicate some tags
which it is unnecessary to apply. (I certainly don't feel like tagging
oneway=no on most everything, for example).

> Cycling is one thing, appropriateness of feet is another, no?

There are basically 3 options:

Imply foot=yes (and several equivalents)
Imply foot=no
No implication

I feel that leaving it with no implication is a bad idea, because
someone wishing to rely on the OSM data for a routing app will need to
have a default for it anyway, either to route foot traffic along a
cycleway, or not.  So OSM might as well indicate which should be
assumed.  Of course, it would be possible to put an additional foot=*
tag on every cycleway, but I think it's better to do this in only half
(or less) of the cases, than to have to do it everywhere.

That leaves us with yes or no.
For the (OSM) definition of a highway=cycleway, it says "mainly or
exclusively for bicycles" -- foot=yes would apply to those which are
"mainly" for bicycles, while foot=no would apply to those which are
"exclusively" for bicycles.

The best assumption is the one which applies in the largest number of
cases.  I think that would be foot=yes.

I know Andy disagrees with this, and I can see his point about foot=yes
being wrong for some countries -- but I think it's better to have a
default so it's easier to make assumptions about routing, than to
require that all cycleways be tagged with an additional foot=* tag.

-Alex Mauer "hawke"

More information about the talk mailing list