[Tagging] cycleway:lane=advisory

Tue Feb 2 13:22:15 UTC 2021

Okay, so there are 485k cycleway=lane and variants and 15k 
cycleway:lane=* and variants.

So for 470k (97%) of all cycle lanes tagged, it is unknown whether they 
are exclusive ("full") lanes or advisory ("dashed") cycle lanes. Both 
exclusive and advisory lanes have been tagged with cycleway=lane because 
(until 2 years ago[1]) there was no way described how to distinguish 
them. That is the de-facto situation.

When we want to record more detail on the cycle lane type and/or we feel 
that the tag is ambiguous, there are only 2 real options how to go about 
this and 1 bogus option.

Two years ago, after the previously linked discussion, I chose to 
document option 1, because it is the best.

Option 1:
------
Use a subtag to define what kind of cycleway=lane it is in more detail.

This is done almost everywhere. type of bollards, type of surveillance, 
type of artwork, of memorial, of recycling, of bench, of doctor, etc 
etc. etc.

This is the least invasive approach. Data consumers who are just 
interested in a lower level of detail, f.e. only process that there is a 
doctor, but not what type of doctor; only process that there is a cycle 
lane but not what type - are not affected.

Option 2:
------
Deprecate cycleway=lane because it is ambiguous and migrate to 
cycleway=exclusive_lane and cycleway=advisory_lane instead.

This has been done on a few occasions in the past, most notably for 
landuse=farm to farmyard or farmland. surface=cobblestone to sett or 
unhewn_cobblestone might count, too.

However, the former was accompanied by a large concerted effort by the 
community. It was doable because farmland and farmyards are very easy to 
tell apart from aerial imagery so it could be done from the home 
computer. All data consumers that supported landuse=farm also had to 
adjust after the break.
For the latter, there are still and likely always will be loads of 
surface=cobblestone around. This is a property one can't ascertain from 
aerial imagery. So, surface=cobblestone just means "either sett or 
unhewn cobblestone". Thus, sett and unhewn cobblestone are a bit of a 
subtype of surface=cobblestone just like building=detached is a subtype 
of building=house which is a subtype of building=residential which is a 
subtype of building=yes.
This implicit subtyping is a little difficult to handle.

Bogus Option:
------
Add a new cycleway tag alongside the current cycleway=lane tag, f.e. 
cycleway=advisory_lane.

By adding this new tag value, it implicitly redefines cycleway=lane to 
mean "only exclusive lanes" and thus an unknown number of these taggings 
would abruptly be wrong. This is why such redefinitions de-facto don't 
happen.
Also, this approach is problem because there is no way to tell whether 
any cycleway=lane has already been resurveyed and adjudged to really be 
an exclusive cycle lane or whether it has not been surveyed yet.

This is the reason why this is not a real option.

[1] https://wiki.openstreetmap.org/wiki/Key:cycleway:lane

Am 31.01.21 um 23:06 schrieb Volker Schmidt:
> OK, let's take this up again.
>
> I think we can live with the advisory lane if we agree that this means cars
> are advised to keep out.
> But, this must be a new value, not a sub:value.
> For the very simple reason that there is no way of adding
> cycleway:lane=exclusive to the existing 500k cycleway=lane and variants
> like cycleway:right|leftboth=lane, in order to distinguish them from the
> new lane-sub-type.
>
> I suspect that many advisory lanes are tagged as traditional "full" lanes.
> On the French Bicycle wiki page they describe the tagging of their
> Chaucidou roads, which have soft or advisory lanes on both sides with
> simple cycleway=lanes tag.
>
> So it will in any case be a national scheme, bout which I am not terribly
> happy.
>
> Volker
>
> Here are the numbers from Taginfo
>
>
> 307 581
>
> *cycleway*<https://taginfo.openstreetmap.org/keys/cycleway>
>
> *lane*<https://taginfo.openstreetmap.org/tags/cycleway=lane>
>
> 124 329
>
> *cycleway*:right<https://taginfo.openstreetmap.org/keys/cycleway%3Aright>
>
> *lane*<https://taginfo.openstreetmap.org/tags/cycleway%3Aright=lane>
>
> 47 113
>
> *cycleway*:left<https://taginfo.openstreetmap.org/keys/cycleway%3Aleft>
>
> *lane*<https://taginfo.openstreetmap.org/tags/cycleway%3Aleft=lane>
>
> 5 310
>
> *cycleway*:both<https://taginfo.openstreetmap.org/keys/cycleway%3Aboth>
>
> *lane*<https://taginfo.openstreetmap.org/tags/cycleway%3Aboth=lane>
>
>
>
>
>
> 1 891
>
> cycleway:right:lane
> <https://taginfo.openstreetmap.org/keys/cycleway%3Aright%3Alane>
>
> advisory
> <https://taginfo.openstreetmap.org/tags/cycleway%3Aright%3Alane=advisory>
>
> 1 606
>
> cycleway:both:lane
> <https://taginfo.openstreetmap.org/keys/cycleway%3Aboth%3Alane>
>
> advisory
> <https://taginfo.openstreetmap.org/tags/cycleway%3Aboth%3Alane=advisory>
>
> 1 414
>
> cycleway:lane<https://taginfo.openstreetmap.org/keys/cycleway%3Alane>
>
> advisory<https://taginfo.openstreetmap.org/tags/cycleway%3Alane=advisory>
>
> 994
>
> cycleway:left:lane
> <https://taginfo.openstreetmap.org/keys/cycleway%3Aleft%3Alane>
>
> advisory
>
> 606
>
> Cycleway:lane
>
> exclusive