[Tagging] What does bicycle=no on a node means?

Mateusz Konieczny matkoniecz at tutanota.com
Thu Oct 22 03:53:18 UTC 2020




21 Oct 2020, 22:00 by tagging at openstreetmap.org:

> On 16/10/2020 09:31, Mateusz Konieczny      wrote:
>
>> Oct 15, 2020, 22:18 by >> tagging at openstreetmap.org>> :
>>
>>>> This recent wiki change by >>>> Emvee <https://wiki.openstreetmap.org/wiki/User:Emvee>>>>>  is in my view not helpful,          or even misleading, as it does discourage a wide-spread          tagging practice (if we like this or not is a different          question, but it's established tagging, and the wiki is          supposed to describe the establsihed methods of tagging)
>>>>
>>>
>>> The change describes what a router does with bicycle=no on a          node, see >>> https://github.com/abrensch/brouter/issues/265
>>>
>>>
>> No, you changed documented meaning of tagging scheme in
>> https://wiki.openstreetmap.org/w/index.php?title=Tag%3Ahighway%3Dcrossing&type=revision&diff=2043653&oldid=2025128
>>
>> OSM Wiki is not describing only tagging that is supported.
>>
>> Note that it is fine to describe tagging as problematic,        unsupported and having a better alternative.
>>
>
> Rereading what was added the text describes exactly what is      problematic namely bicycle=no in the context of routing. I did not      add that context but that is something I can do. 
>
>
> Adding that mapping the crossing from curb to curb as separate      osm way with the correct access tags is a better alternative is a      good idea.
>
>
>>>
>>> Already discussed elsewhere but having routers ignore          bicycle=no in combination with highway=crossing means that it          is more or less useless as routers are they main data          consumers while at the same time crossing data is far from          being complete.
>>>
>>>
>> Any tagging scheme is for some period unsupported, this does        not make it useless.
>>
> If data is not used and will not be used in the foreseeable future I    call it useless.
>> And any widely used tagging scheme can be described. As        obvious from this discussion meaning
>> of this bicycle=no is clear so I will revert your edits on        this page
>>
>
> I do not see how you came to this conclusion, but as I noted on      the Talk page I have no problem with reverting for now but think      it should be reverted further to point before bicycle=no/yes was      added.
>
>
Why?
>
> Instead of reverting you could have chosen for the changes I did      point out above.
>
>
>>>
>>> My take is that it is not a wide-spread tagging practice and          it does not add new information as weather it is a pedestrian          issue can be deduced from the connecting ways.
>>>
>>>
>> Not in cases where 
>> (1) highway=cycleway is crossing road where cyclists are        obligated to dismount
>> (2) highway=footway with bicycle=yes/designated is crossing        road where cyclists
>> are obligated to dismount
>>
> Can be covered by mapping the crossing, curb to curb as separate osm    way. A bit more effort but more precise.
>
Yes. But the entire thread was started due
to wiki edit redefining already tagged data.

I am open to describing way splitting as
preferable, but not to redefining existing tagging.
>  
>
>>
>> (3)pedestrian only crossing is tagged on road having cycleway        on both sides 
>> (tagged as cycleway:lef/cycleway:right/cycleway:both) 
>> (or where such road has cycleway at one side, is joined by        separately mapped 
>> cycleway from other side and there is crossing there, but
>> cyclists must dismount)
>>
>
> There is no need to tag this type of "solitary" crossing for      routing purposes, a router will never want to make a turn half way      the road. 
>
>
But given road may be obstacle to be crossed
with user desiring to get to other side
>
> So these "solitary" crossings are useless in routing      context while routers do have problems with bicycle=no/dismount on      a node.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tagging/attachments/20201022/563961a5/attachment.htm>


More information about the Tagging mailing list