Classification of Cycle Paths

Thanks Masimaster and Rayquaza for the examples. Althought Rayquaza’s examples all seem te be tracks I still get the picture. I’ve aslo seen that Wolmatinger has added some FIXME’s on these tracks so maybe a local mapper could fix this.

I did not know this. This will make my cycling trips in Germany a bit faster :wink:

I think you are right. I also think this is what Wolmatiger means. I guess I was thinking of roads next to compulsary cycleway. If those roads would get a bicycle=avoid (tag to be discussed) for those parts were a policeman could give you a fine you could add a bicycle=yes just for the parts of the road were it is OK to use a bike. You could still say it is redundant but maybe you can compare it to the next situation. In NL many cycleways are oneway and are also tagged as oneway=yes. For this reason some mappers use a oneway=no for cycleways that can be used in 2 directions. Just to make sure no one accidentely maps these as oneway. This oneway=no is also redundant but prevents athoher mappers from accedentaly mapping these as oneway. Sometimes it is just a little part of the cycleway that is 2-way but if it was mapped as oneway it would be wrong and also affect bicycle routing.

So what do you think of the next conclusion?

  1. All roads with a “cycling forbidden” sign get a bicyle=no
  2. All (parts) of a road next to a compulsary cycleway (blue sign) were you could get a fine by a policeman when riding a normal bicycle get a bicyle=avoid (tag to be discussed)

This would improve bicycle routing I think. What do you think? Could this work in Germany or am I fighting windmills :wink: ?

This would be a “non-local” part of a routers cost function and all the tagging you need for that is already there, but this is nearly impossible to implement and not a single router is doing stuff like that.

This is why I share the opinion that an extension of the tagging scheme is neccesary to allow correct routing via local cost functions.

No extension needed here. Remind that a road with cycling not allowed does not neccesarily need bicyle=no. motorroad=yes implies bicyle=no, and check against bicyle=dismount.

+1 for that

The windmills you are fighting is the “All (parts)” remark. It should be made clear that this is not a strict access rule and that it is NOT necceary to fiddle with the “edge-issues” around crossings. If a router has no other choice than routing through a 5 Meter section of bicycle=avoid than he should just do it.

Thanks. It’s difficult to be clear and 100% correct. I guess this scheme could be the guide for when to add bicycle=no.

I’m glad you like this. I guess it would greatly improve your “routing engine” for cycling (normal bike) and give the oppertunity for routing for special bikes/ groups of race cyclists

+1 although I think some of this issues can be solved with a little “better” mapping.

It looks like response for this proposal is not overwelming. Reasons I can think of are:

  1. It could be that many agree/disagree but don’t say so
  2. The english language of this thread keeps some away from responding/reading
  3. Some don’t like me as a foreigner interfering in German tagging

I thinks it’s best if I try this proposal in the Dutch forum with many active cyclists. In NL it is not allowed (yet?) to cycle these roads with a group of race cyclists but… discussions about this are alive. So maybe if the law changes the proposal will get more support :wink:

If someone wants to start a thread about this proposal (in german this time)… be my guest.

Only the one where I wrote so – the others are signed with a cycleway-Sign (but actually one of them is a highway=track from the other side).
@Wolmatinger: Bis auf den =yes-Hinweis sind die fixme= unpassend: Das eine sieht real wie ein normaler Gehsteig aus und das andere ist (iirc) (teilweise) wie eine normale Strasse (=gar nicht) ausgeschildert. I am a local mapper there :wink:

I think 1) is common in OSM. (motorways, or barriers who you can’t cycling gets a bicycle=no too)
And +1 to No. 2), including “tag to be discussed”!

Furthermore I think we have to discuss about the tags for the cycleway:

  • designated: this is a tag for a way with a sign?! It tells nothing about the strict law i.e. about cycleways. So some German mapper introduce bicycle=official for the strict access.
  • yes: this tag could to be used for cycleway where cycling is allowed. But the problem with tis is: many mappers tag some ways (mostly on path) with bicycle=yes, where cycling is possible. a part of this problem are the editors with the select menu… I think cycling=yes, walking=yes and riding=yes sounds better for only the possibility.

If we use designated for the strict law, we get a problem with designated for other ways which have a sign. And if we use bicycle=yes for signs which tell “cycling is allowed”, there is the problem i explained at “- yes”.
I think it is important if we want tag only one sign with one tag/one tag correlate to one sign, which may be helpful for OSM.

OK thanks.
The first example is tagged as a highway = track. (clicking way will give OSM tags). But according to this scheme (though still proposal) would not requiere tags like foot=yes, bicycle=yes. In that way I can understand Wolmatigers remark. From the taggings I would not have guessed it was a compulsary cycleway. Thats the reason Mapnik renders it as a track (grade1) and not a cycleway.

You are right about the second example. It is not a track but a highway= footway. As you can see Mapnik does not render these as a cycleway but as a footway. I thought combined cycleway/ footway that are tagged as a highway=footway would get a bicycle=designated in Germany. This one has a “bicycle=official”. I guess that is why Mapnik does not render this as cycleway but as footway.

Masimaster… we are brothers in arms ;). I aggree with what you say on both the issues of ways where cycling is (most of the time) forbidden and on tagging of cycleways. I also think it is important to keep these 2 issues apart not make discussions more complicated then it already is.

I have started a thread about the type1 and type2 roads where cycling is (most of the time) forbidden. A google translate version (NL>DE) is here but I do not know if this makes any sense so do not believe every word you read :wink:

Long story short. I think most understand that there is a need for a new tag in Gemany (edit: not only Germany ofcourse) to keep these 2 ways apart which would help routing for cyclist. In NL there is not such a great need because it is costum to also tag type2 roads with a bicycle=no. This is the reason why routing for normal bikes in NL works better then in DE. But if in other countries (eg Germany) this new tag would be accepted and commenly used I am sure more countries (edit: including NL) will follow. Also routers/renderers would follow.

I’ve been thinking about the name for this tag. What about “bicycle=no_unless” . The “unless” indicates that there are exceptions. These exceptions vary from country to country but I am sure routers/renderers can deal with this. Still a lot better then tags like “bicycle_longer_then_x_wider_then_y =yes” or “group_of_cyclist_more_then_x_bicycles=yes” :wink:

That’s the point: It is signed with a cycleway-Sign but since it leads away from the road it isn’t compulsory (but you can’t see this from where you would have to change to it). If you approach this intersection from this “Cycleway” you only see a Cycleway-End-Sign, no Cycleway-Begin-Sign.

As I wrote above in german it actually looks like a normal sidewalk, so the rendering isn’t that wrong. I’m not sure where the bicycle=official comes from, maybe I found that somewhere in the Wiki or on an other Way.

About what Google Translate produced with the NL-Thread: It’s hardly readable. But here a few annotations:

  • Reply 4: Isn’t exactly that the advantage of OSM? That we also have data for minorities? Remember: Only a few Percent of our potential users would use things like wheelchair=* (which I find totaly useless, but that’s another topic).

  • Reply 9: If I understood correctly (“You should earn a tenth of your teachers”?) I have to aggree with that.

  • Reply 10: It is controversial if you would be allowed to use the road if you think the cycleway is too bad. If a policeman stops you using the road you’re “guilty” (even if there’s a cycleway-damage-sign) and if something happens because you used the cycleway you’re “guilty” too…
    And i.e. parking cars are usualy not signed, but they still allow you to use the road between two entrys to the cycleway (which often means between at least two intersections).

Slacknas of the responsible adminstration. A problem in Germany.

e.g.

  • sign 240 instead of sign 250 + Rad frei or sign 260
  • sign 239 is replaced through sign 254
  • missing bicycle=yes sign on ways with sign 250 + vehicle/access=destination/agriculture which are signed as bicycle-route

I think we can understand the translation (the translation is so bad, because it works: NL → EN → DE. often it makes more sense, read the english translation, even you cant understand each fourth word :slight_smile: )

I think the bicycle=no works for NL, because there are very few cycleways (and most >90% ?? are good), so there is no reason to cycling on the road.
In DE most cycleways (>80%) are very bad and often car-drivers use them for parking and pedestrian for walking without looking for (bicycle-)traffic. So often it is allowed to cycle on the road.

The example with the green and red arrows: in Germany you may cross the crossing over the road, because there is no parallel cycleway (in a 5m distance). And because it is not possible to sort you on the right lane for cycling straight, it is allowed to switch to the road a intersection before. (It only works, if you know the crossing.)

In Austria riders with a racing bike on a training-tour, they are allowed to drive on the road! only a “bicycle=no-sign” can stop them (as well as motorway etc.)

I’m not really happy with no_unless, because the double negation (no & un) and the two words. But i don’t have a better idea.

I completely agree! The administration in Germany do nothing for cyclists! Also the ways with are not parallel to the road, here the official bicycle sign are wrong. (I think this is because we don’t have a no sign like in france: http://imageshack.us/photo/my-images/265/andererradweg.png or the “fietspad-sign” in NL). But this is a problem in Germany and out of OSM.

Her I go again. When I wrote “compulsary cycleway” I meant a cycleway with a bleu sign/bicycle. In NL is it common to call these compulsary but in DE it is not always so… Communication… difficult as always :wink:

correct… Danke :wink:

I don’t know if I understand correctly but in NL it is not allowed to park a car on a cycleway so here’s an other difference between the 2 countries.

I understand this also goes for some “Bundesländer”. I heard from a Dutchman living close to the Niedersachsen/Holland border that is is also allowed there.

On the Dutch forum Ligfietser came up with the “bicycle=use_cycleway” . The best so far. It is difficult (if not impossible) to find a tag to discribe all the exceptions so this one seems pretty good to me.

In DE it’s also not allowed.

It is also prohibited in germany, but still cycleways get “unusable” this way…

This might be allowed in germany, but I’m not sure about this (here we have one of those “and maybe even more exceptions”). If so it is likely the same for whole germany.

OT: A “Bundesland” is called “federal state” in english :wink:

I think all no Bundesland has special rights, so it should belong to the whole country. I can’t find any right that tell that racing-cyclists may ignore the cycleways in germany.

+1 “bicycle=use_cycleway” is the best so far!

But i think we don’t need describe all exceptions in the tag/tagging. It should be enough to explain it for different countries here:
http://wiki.openstreetmap.org/wiki/OSM_tags_for_routing/Access-Restrictions

Yesterday I encountered a routing glitch on a long-distance trip using BRouter/fastbike.

A section on B8 in Bad Camberg (Germany) is tagged as a combination not accepted by brouter/fastbike:

  • For the road: highway=primary, bicycle=no
  • For both cycleways: highway=path, oneway=yes, bicycle=desiganted, foot=designated, segregated=yes

so I got a painful detour up the hill.

Since tagging for the router is evil and because I need to get brouter to work on the existing map material, I now changed the fastbike-profile to accept that situation, I did not change the map.

But the question is: is the bicycle=use_cycleway proposal in a state that it would be allowed to adapt the tagging? And what about the oneways on the cýcleways ?

Hi,
wieso geht BRouter denn nicht über die Radwege?

oneway: Ja, das sollte für Radler natürlich beachtet werden.

Zum proposal: Das wurde leider abgelehnt.

http://wiki.openstreetmap.org/wiki/Proposed_features/Bicycle_use_cycleway

Chris

The proposal is indeed rejected but we’re working on a new version. That may take some weeks (i hope). Although I prefer an accepted wiki page nobody will forbid you to use the tag. In Bremen there are many tagged. This may be a nice test area for your router. In the new version of the proposal we want to use some routing examples. Maybe you can help us making clear what the advantage for routing can be.

Es betraf nur das ‘fastbike’ profile, und da hatte ich gesagt, ein “ungraded path” wird nicht akzeptiert, weil das im Zweifel auch ein Matchweg ist. Jetzt taggen die Mapper aber oft für kombinierte Rad/Fusswege highway=path, weil es ja weder ‘cycleway’ noch ‘footway’ ist. Ich hab’s jetzt so geändert, dass das ‘bicycle=designated’ den Weg dann doch wieder qualifiziert.

Schwierig wird das für die Zweiräder, die die Radwege nicht benutzen dürfen. Beim Moped-Profil verzichte ich ganz auf die bike-access tags und arbeite nur mit den impliziten Berechtigungen. Ob das für S-Pedelecs so auch funktioniert weiss ich nicht. Heisst ja, dass Fahrrad-Verbotsschilder ignoriert werden, nur das Autostrassen-Schild (motorroad=yes) zählt.