[Tagging] Sidewalks and cycleways as tags vs as extra lines

"Christian Müller" cmue81 at gmx.de
Wed Jan 17 11:39:13 UTC 2018


> Gesendet: Dienstag, 16. Januar 2018 um 22:13 Uhr
> Von: "Fernando Trebien" <fernando.trebien at gmail.com>
> An: "Tag discussion, strategy and related tools" <tagging at openstreetmap.org>
> Betreff: Re: [Tagging] Sidewalks and cycleways as tags vs as extra lines
>
> For simplicity, I think it should be no big problem changing the wiki
> to suggest using cycleway=separate, instead of suggesting not adding
> the cycleway=* tag.

I partly agree, but the word is not a good choice.  It is easily confused
with "separate track" (as opposed to lane) _or_ the semantic of segregated=*
which informs whether the cycleway is "separate" or shared with/on the side-
walk.  For the same reason it is not a good choice to use as a sidewalk=*
value (despite the fact that it already is).

We should try to improve this in a coherent manner for all alongside features,
Ideally we should start a proposal to get thoughts and comments from a broader
audience on this and gain acceptance for any proposed change.

It is never a good idea to mix different classes of information within the
same tag.  Hence, it also needs debate, whether "separate" (as in separate
element / separate osm object) should be a sidewalk=* value recommendation
in the wiki.

As far as I can tell, we never differentiate between sidewalk lanes and tracks,
which suggests that lanes are solely marked for cyclists, never for pedestrians.
(But is this really true, worldwide?)

Here is a revised overview of what may be tagged on the motorized way,
if a parallel way exists, in decreasing importance

- to which side of the road sidewalk / cycleway infrastructure exists
::values: left, right, both, none

- the type of (sidewalk /) cycleway infrastructure
::values: lane, track
This is assumed to be "track" for sidewalks, i.e. we expect any sidewalk to
be structurally divided from the road surface (by definition of "sidewalk").

- direction of (sidewalk /) cycleway traffic flow
::values: forward, backward, both  (wrt direction of the osm way primitive)
This is assumed to be "both" for sidewalks, always.

- whether the sidewalk / cycleway is mapped separatly
::values: yes, no
It indicates a weak relationship (a relation primitive is not used, but it
speaks about another primitive in the data, rather than a real world aspect).


All of these tags may be useful to a renderer that wants to do fancy things
with the encasing lines on low level zoom scales to encode details that would
otherwise be hidden or disturb map appearance on these scales.  All of these
tags are redundant, because they may also be computed by looking at the paral-
lel way running along (under the premise it exists).

Note that things could be simplified by pulling apart cycleway lanes and tracks,
because since lanes are predominantly structurally undivided from the road sur-
face, they are expected to be tagged/described on the motorized way and never as
a parallel way (according to established practice to not map individual lanes).

Say you use cycletrack and cyclelane tags instead of cycleway, then it is easier
to construct a coherence between sidewalk=* and cycletrack=* (than it would be
between sidewalk=* and cycleway=*).


Greetings
cm



More information about the Tagging mailing list