[Tagging] Tagging cycleways to distinguish them from combined cycle and footways

[Jens Glad Balchen]

Hi Mateusz. Thank you for responding.

On 03.08.2022 09:16, Mateusz Konieczny via Tagging wrote:
> Aug 2, 2022, 23:42 by balchen at saint-etienne.no:
>
>     Hello everyone.
>
>     I'm in the process of creating a cycle road map of the Stavanger
>     region in Norway. One of the desirable outcomes of that process is
>     to be able to show the various forms of cycle roads that we have,
>     and to visually illustrate how much of each we have, and where we
>     have them. To accomplish this, the OSM data in the region has
>     needed a lot of cleaning up to accurately reflect what is on the
>     ground, since tagging practices appear to have changed over time,
>     and have been (and probably still are) implemented inconsistently
>     between contributors.
>
> Good luck with the project! If you produce online-viewable map - feel 
> free to email
> also me!

The default map from the tile server is viewable here: 
https://sykkelkart.rogalandsyklistforening.no/
Move to the Stavanger area to see what a fully cleaned map looks like.
Tiles are rendered on the fly.
Due to react-leaflet having made itself incompatible with 
create-react-app, I've not yet set up a proper map viewer.

> Note that you can produce (shown on the main map or in additional one) 
> display
> of where OSM data is dubious or missing important tags and share it 
> with a
> community.

For zoom level >= 17, I include highway=cycleway and highway=footway 
that do not match the quality requirements, wether it's from a lack of 
tagging or from tagging that indicates too low quality (e.g. 
surface=fine_gravel is normally excluded, no surface tagging at all is 
normally excluded, and highway=footway is normally excluded, but they 
are all included at zoom >= 17).

> You can also ask authors of QA tools to such as JOSM validator to complain
> in clearly dubious cases not reported already.

That would be very useful, since it seems at least the iD editor at the 
moment enforces a different taggging standard from the documented 
tagging standard at https://wiki.openstreetmap.org/wiki/No:Map_Features.

> We share common legal definitions of cycle roads with much of Europe, 
> in that we have cycleways designated for cycling, that are legally 
> accessible to pedestrians, and combined cycle and footways that are 
> designated for both groups. Cycleways may or may not have a sidewalk 
> for pedestrians, and may or may not have separated lanes.
> This seems to not match later claim that
> "cycleway with no sidewalk is tagged highway=cycleway + 
> foot=no/discouraged"

These two claims are about two completely different things. The first 
claim is about the legal status, whereas the second claim is about the 
tagging standard at https://wiki.openstreetmap.org/wiki/No:Map_Features.
*This is the very heart of the issue that I am raising -- that the 
tagging standard **/does not reflect the legal status./*

> Is standalone cycleway without footway part accessible to pedestrians 
> or not?

It is, by law. https://lovdata.no/dokument/SF/forskrift/1986-03-21-747, 
§ 19, auto-translated:

"Pedestrians must use the footpath, pavement or shoulder of the road. Is 
it not reasonable because of the speed etc. or possible to do this, 
pedestrians can use cycle paths, cycle lanes or carriageways."

>     The current tagging standard (per
>     https://wiki.openstreetmap.org/wiki/No:Map_Features) is to tag a
>     combined cycle and footway with highway=cycleway +
>     foot=designated. A cycleway with a sidewalk is tagged
>     highway=cycleway + foot=designated + segregated=yes +
>     sidewalk=left/right. A cycleway with no sidewalk is tagged
>     highway=cycleway + foot=no/discouraged. There's a parenthesis
>     saying foot=no/discouraged applies if the cycleway is not intended
>     for pedestrians.
>
>     The strict interpretation of this standard is that
>     highway=cycleway by itself denotes a cycleway, designated for
>     cycling, where pedestrians have legal right of access. The
>     feedback from the few participating Norwegian OSM forum users,
>     however, is that any highway=cycleway should be regarded as a
>     combined cycle and footway,with reference to how most people and
>     OSM contributors are not aware of the differences between the
>     various types of road. This approach requires foot=no/discouraged
>     for a road to be a regarded as a cycleway proper.
>
> I am unfamiliar with Norway and whether their foot=no/discouraged on 
> cycleways
> without footway part is a good tagging or not.
>
> Maybe it matches actual legal rules and pedestrians are actually 
> discouraged/banned
> there - or maybe it is some kind of
> https://wiki.openstreetmap.org/w/index.php?title=Mistagging_for_the_renderer

A probable contributing cause is "mistagging for the router".

Pedestrians are generally allowed on any road, with some mild 
constraints. Cyclists are generally allowed on any road and on 
sidewalks, with some constraints on speed. This is a rather uniquely 
Norwegian situation. The observable tagging practice in Norway seems to 
have been to make sure that everyone is allowed everywhere so that 
routing can work even if routers do not implement Norwegian law.

Another probable contributing cause is that fact that most people, even 
road planners, don't know a cycleway from a combined cycleway and footway.

All these tings combined have caused a "proper" cycleway to become an 
exception that can no longer be accurately described in OSM, so practice 
is to tag them with foot=no/discouraged even if this is wrong according 
to the law.

> But based on this description it appears that maybe highway=cycleway 
> without
> further info should be treated as incomplete tagging?

If that is the conclusion, which tag would we add to complete it? 
foot=no/discouraged is wrong by law. What else?

>     The strict interpretation of this standard is that
>     highway=cycleway by itself denotes a cycleway, designated for
>     cycling, where pedestrians have legal right of access.
>
> I do not see it from your description.

https://wiki.openstreetmap.org/wiki/No:Map_Features says to tag a 
cycleway with highway=cycleway, and optionally add foot=no/discouraged. 
The strict interpretation is that a cycleway is defined by 
highway=cycleway on its own, since foot=no/discouraged is optional. The 
fact that pedestrians have legal right of access on any cycleway where 
not explicitly prohibited, follows from the law.

>
>     _Tagging foot=no
>
>     _The highway code explicitly allows pedestrians the use of
>     cycleways and carriageways when or where they find that it is not
>     possible, practical, or safe to use a different road or the road's
>     shoulder. It follows that foot=no can only be (correctly) applied
>     when there is a sign prohibiting entry for pedestrians. foot=no
>     can never generally be applied to cycleways as an consequence of a
>     road being a cycleway.
>
> Based on info provided here it makes sense.
> Maybe it also makes sense if there is a cycleway and footway along 
> each other,
> mapped as  separate lines,
> like for example on this Polish footway + cycleway
> https://commons.wikimedia.org/wiki/File:Bike_road_-_panoramio_(1).jpg
> https://commons.wikimedia.org/wiki/File:2019_Warszawa_aleje_Jerozolimskie,_chodnik_i_droga_rowerowa.jpg

The law (as quoted above) allows pedestrians to use their own judgement 
on when and where they find it reasonable or possible to use the 
footway, sidewalk or shoulder. As such, it cannot principally and 
generally be tagged with foot=no even if there is an adjacent footway, 
unless there is a sign.


> But if one chooses to interpret the highway code that way, pedestrians 
> are equally discouraged from using carriageways. It follows that this 
> interpretation of pedestrians being discouraged from using a road is 
> derived from the type of road.
>
>
>     Tagging foot=discouraged on a highway=cycleway in this scenario
>     would be optional, explicit, and redundant, and equally so tagging
>     foot=discouraged on every
>     highway=trunk/primary/secondary/residential/service/unclassified.
>
>     Note that in either case, using the road's (hard or soft) shoulder
>     is /always/ /explicitly allowed/ -- the only discouraging one can
>     possibly interpret from the highway code is from the use of the
>     cycleway/carriageway itself.
>
> I am not really understanding this part.

The carriageway (or cycleway) is defined as the part of the road used 
for regular vehicle traffic. Typically this area is marked by solid or 
dashed white lines on the edges of the road. Anything outside of those 
lines is the road's shoulder and is /not/ a part of the 
carriageway/cycleway. Norwegian law explicitly allows pedestrians to 
/always use the shoulder on any road //unless explicitly prohibited/.

Driving theory questions: Study the picture. Which letter represents the 
hard shoulder?

>     _Tagging foot=no/discouraged if the cycleway is not intended for
>     pedestrians
>     _
>     _
>     _This phrase makes sense if there is a sign prohibiting walking,
>     in which case the only correct tag is foot=no. In all other
>     situations, pedestrians are explicitly allowed access by law. Any
>     intentions of planners or officials are neither observable nor
>     verifiable.
>
> also, foot= is for legal access - not for whether access should be 
> present there or
> what designed intended

Exactly. And the legal status is clearly outlined above. Pedestrians are 
allowed, according to their judgement of the specific circumstances at 
the time or place, on any cycleway or carriageway, and they are by 
default always allowed to the road's shoulder, where not explicitly 
prohibited.

>     _foot=discouraged is required to define a cycleway in OSM_
>
>     This logic completely reverses the causality of the most
>     open-minded interpretation of the highway code. How can a cycleway
>     be defined by foot=discouraged when foot=discouraged -- at best --
>     follows from the road being a cycleway or a carriageway?
>
>     _Tagging a cycleway proper in OSM?
>
>     _The intuitive and logical representation of a cycleway proper
>     would be highway=cycleway. If the Norwegian OSM community cannot
>     agree on this being the case, how could we tag a cycleway in a
>     manner that is logical, consistent, and accurate (ref the above)?
>
> Sadly, I am not really understanding this part.

If you have a road that is signed as a cycleway (with no pedestrian 
sidewalk), how would we tag that in a way that allows us to later 
recognise from the tagging that "this is a signed cycleway with no 
sidewalk", and not "this is a signed combined cycleway and footway". 
Tagging with foot=no is not correct, neither is foot=discouraged, since 
pedestrians are legally allowed to use it (due to the already quoted law).

>     Bonus question:
>
>     _Tagging foot=designated on a cycleway with a sidewalk_
>
>     This seems to principally be the same as tagging foot=designated
>     on any highway=* with a sidewalk. It seems weird, and redundant,
>     but probably not harmful, so long as no further meaning is
>     attributed to or derived from the tagging?
>
> What you mean by cycleway with sidewalk? If something like
> https://commons.wikimedia.org/wiki/File:Witosa_bike_2.jpg
> then foot=designated segregated=yes is an useful tagging.

No, I do not mean like that. In this case, the pedestrian is on the sign 
and I assume Polish law defines the meaning of that sign. In Norway, we 
do not have that particular sign. We only have the one for combined 
cycleway and footway, the one for cycleway, and the one for footway.

sf-20051007-1219-522-01.gif sf-20051007-1219-520-01.gif 
sf-20051007-1219-518-01.gif

This is how a cycleway with a sidewalk is signed in Norway:

https://miljopakken.no/wp-content/uploads/2019/11/Stiklestadvn-m-syklist.jpg

Road painting does not carry any legal status in Norway and is only 
there to help guide road users to the appropriate

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tagging/attachments/20220803/e2bdb0a0/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: lGfHmTWa8v4B7uIL.png
Type: image/png
Size: 3456 bytes
Desc: not available
URL: <http://lists.openstreetmap.org/pipermail/tagging/attachments/20220803/e2bdb0a0/attachment-0001.png>