Add support for "cycleway:both" tags

[FFMbyBicycle]

Currently we can only check for "cycleway=both", but not the way more detailed "cycleway:both=*" tagging:

https://github.com/abrensch/brouter/blob/master/misc/profiles2/lookups.dat#L313-L333

There are more than half a million used tags and it would be nice if we could use them: https://taginfo.openstreetmap.org/keys/cycleway%253Aboth

If somebody tells me how to add new tags, I will push some PR for this and i.e. #398.

[EssBee59]

I was interested to estimate the impact of the "cycleway:both" tag, here the results for Hessen/Germany (got with overpass-turbo):

total occurences: 18.547

number of occurences by value:
=lane 671
=no 17673
=none 5
=separate 41
=shared_lane 54
=track 99
=yes 4

Yes.it would be nice if we could use the tag in scripts (by nearly 900 highways the existing cycleways are not visible currently)

Nearly 12.000 highway are tagged wit the old/standard "cycleway=" , 4.000 of them with the "no" value.

My prefered bike is the "fastbike" (I created my own profile for very-low traffic), especially for this kind of profile it is interesting to see all the highways having a cycleway

A further tag exists,"cycleway:lane", but it is only used 59 times in the same region (value is quite all =advisory), not sure it make sense to add it too..

[afischerdev]

I agree, this would a nice feature.
The bicycle streets exploding in my area as well.

Let us have a look if there are some other aspects for lookups.dat (like #398 or #402) and then start an update.

I could collect the lookups data and make a suggestion.
Is there someone who like to do the update on the standard profiles?

[polyscias]

I am not so sure I would like to have an update of the profiles that somehow prefer certain cycleway:both values. I have seen figures telling that roads with cycleways (apart from tracks) are no safer than the simular way without cycleway tracks. If anything is important it is the width.

Let us have a look if there are some other aspects for lookups.dat (like #398 or #402) and then start an update.

Would be good to start a new issue for this and gather there all information I think.
I like to review the current lookup.dat values, I think some can be removed, but let me provide data for that.

[EssBee59]

Hello Polyscias,
The differences between "cycleway=" and "cycleway:booth=" are very very low...
Do you consider "cycleway=" currently in your biking scripts?
If roads with "cycleway" are not safer than similar way without cycleway, then I suspect a mapping error / issue.

Not every biker is using the cycleways (a.e. bikers in group at week end) but others (bikers riding "alone" middle in the week on roads with trucks!) prefer safe lanes or tracks on the side of the road..
If the change is accepted, my intention is to consider "cyclewa:both" exactly as "cycleway"

Starting a new issue to gather all informations is a good idea!!!
I could work on that next week, but I am not sure to be able to document exactly all change requests in lookup:
Perhaps is your thought, every "requester" should make a contribution to this generic issue?

[polyscias]

Yes, the differences between cycleway= and cycleway:both= are small, so agree, if cycleway= is taken into account, so should cycleway:both= be taken into account.

The profiles I use use only cycleway=opposite and it's variants (outdated) as alias for oneway:bicycle=no and I have some code that makes the profile handle cycleway=track similar as highway=cycleway.

If roads with "cycleway" are not safer than similar way without cycleway, then I suspect a mapping error / issue.

Do you think a cycle lane of 50 cm with cars parked on the right is safer then no cycle lane? My point is that OSM is mapping all kind of attributes that are, at least for me, not important, what I think is important is width, if cars are parked parallel and the speed of car traffic.

I think a generic issue can be used to summarize things, for longer discussion it is likely better to use a separate topic.