Talk:Proposed features/use sideway

From OpenStreetMap Wiki
Jump to: navigation, search

PLEASE SEE NEW PROPOSAL : bicycle=use_sidepath. Same thing, different name.

Just one advice

To increase your chances, you should write a shorter proposal. You repeat the same arguments 2, 3 or even more times. The header contains "bicycle(/foot)=use_sideway". Either make it properly for foot : ""bicycle=use_sideway, foot=use_sideway"; or remove any mentions of "foot" completed and concentrate your efforts on bicycle (it will be easier to create the "foot" proposal if the "bicycle" one is adopted) --Pieren (talk) 09:48, 17 March 2014 (UTC)

Thanks. We agree. We've changed things in the proposal--PeeWee32 (talk) 07:30, 21 March 2014 (UTC)

bicycle=no / bicycle=destination

Even if bicycle is not formally forbidden by a road sign, you should explain why your tag is better than "bicycle=destination" which is corresponding to your "like "destination, limited, restricted".". I guess it is the main argument justifying this tag. -- Pieren (talk) 09:48, 17 March 2014 (UTC)

I am coming back to respond directly to this point, since it may not be obvious that in the Netherlands (in any case) bicycle=destination is NOT a valid representation of the law and regulations EXCEPT in a few very specific circumstances. So the reason not to use bicycle=destination on a road in the Netherlands (except as noted below for access) is because it is incorrect. Perhaps you can tell us how this works in France? Frankl2009 (talk) 22:19, 3 April 2014 (UTC)
Although I agree with this comment (and the previous one on making the proposal simpler and less repetitive), I cannot think of a real world situation in the Netherlands in which bicycle=destination on a road with a parallel compulsory cycleway would be valid tagging. After some discussion on the Dutch tagging list, we do have examples of situations where only a bicycle=destination will allow correct routing (otherwise houses would be orphaned in routing). However these are quite unusual. The Dutch cycleways are well-developed and will allow one to cycle in any valid (i.e. legal) direction. I would appreciate a real-world example which shows why this is needed. Frankl2009 (talk) 10:08, 17 March 2014 (UTC)
I was also hoping that bicycle=destination argument&answer would have been listed in the "reasons for opposing" because it was one of the alternative suggestions to the previous proposal. I'm surprised though that it was even proposed back then as it would be very major misuse of =destination "for router" and would overload its meaning making bicycle=destination entirely useless tag in the end. -- Ij (talk) 20:00, 18 March 2014 (UTC)
PeeWee32 answered my question elsewhere, confirming that also in his opinion we do not need bicycle=destination in our jurisdiction (the Netherlands). In fact, there ARE some (fairly unusual) situations where it is valid and necessary to reach certain locations. TheAnother argument for using this tag is said to be that in some jurisdictions (e.g. Germany) the law/regulations would allow a cyclist to leave the cyclepath under certain conditions, such as snow which has not been cleared from the cyclepath or a very poor (potentially dangerous) surface due to the action of tree roots under the path. AFAIK this is not regulated in Dutch law, and it is not an accepted practice; therefore we would not need to use bicycle=destination at all. I do believe it confuses the discussion unnecessarily. Frankl2009 (talk) 16:04, 22 March 2014 (UTC)
Bicycle=no/destination can't work for 2 reasons:
  • First, there are no "No" or "Destinatin" sign. The same tag should not cover different things.
  • Second, there are several reasons in different countries to use the road instrad the compulsory cycleway, which have nothing to do with destination: The cycleway is damaged/not save/blocked by lowed kurbs/cerbs, parking cars, dooring-zone, dustbins, dirt, broken glass pieces, holes, root-bumps, snow or much wet slithery leafs, traffic signs or street lamps in the right of way, ... (Maybe non-cyclists can't imagine this, but it is common on cycleways in many countries.)
I'm not sure that we should put the discus to this odd value-possibilities. --MasiMaster (talk) 23:04, 24 March 2014 (UTC)
I do not understand this comment. The only case we can currently find in the Netherlands for bicycle=destination is where routing to houses would otherwise be prevented. (So, it is a valid option.) However, there is NO general rule in the Netherlands which allows cyclists to ignore compulsory cycleways. In this respect, it is maybe different from Germany or other jurisdictions. To be clear: you do NOT have the right in the Netherlands to cycle on the road if the compulsory cyclepath is in your opinion unsuitable or dangerous. (Whether you would be prosecuted for such behaviour is a completely different question; prosecutorial flexibility also differs between different jurisdictions.)
Also, if at the start of a road there is a compulsory cyclepath, you have to follow it, even if you know that later on it heads away from the direction you want to travel in. (In such a case, you will need to find a different route to go where you want. You cannot just ignore the cyclepath and head down that busy road. We also found an example of this in the Netherlands.)Frankl2009 (talk) 09:53, 2 April 2014 (UTC)
You prefer bicycle=destination at the road? Why you want to use the same tag for different situations/signs? Especially the Dutch should know the this is not a good solution. They have exactly the same problem with the bicycle=no, for both the bicycle-forbidden-sign and the compulsory cycleway. But there are differences! Even there is no difference, do you propose a compatible tag only for the Netherlands or also which fits worldwide? --MasiMaster (talk) 16:39, 3 April 2014 (UTC)
I am still not sure I understand your point. On the Dutch forum we looked for an example where bicycle=destination on the road would make sense (my own initial position was that we did not need bicycle=destination on a road at all; this proved to be incorrect). We have an example where there is a compulsory cycle path on one side of the road and no way to get to houses on the opposite site of that road unless you allow bicycles to cycle along the road. One solution is to leave the road untagged (since no tag would mean there is no restriction), but for consistency with the other parts of the road which would be tagged bicycle=<use sideway or some other tag> it is probably better to add the bicycle=destination tag (avoids people "completing" the tagging incorrectly). There will not be very many such cases (the regulations in the Netherlands do not allow cyclists to leave the cycle path, except for certain 'exceptional' cycles). Note that the only reason to cycle on the road in this case is to get to a house along that small stretch of road; if you are cycling through the street to get somewhere else, you would be expected to use the cycle path. Frankl2009 (talk) 21:54, 3 April 2014 (UTC)

Highway (i.e. tertiary)

Not all roads in the Netherlands which are secondary roads have explicit signage forbidding bicycles (and mopeds of various kinds) when there is a parallel compulsory cyclepath. The general rule is that bicycles and mofas (and mopeds, if the relevant sign is shown) MUST use the cyclepath and not the road. If the intention is to name "tertiary" as an example, this should read "e.g. tertiary". As written currently, it defines the highways in question as tertiary highways, which may not be the case.Frankl2009 (talk) 09:56, 2 April 2014 (UTC)

Thanks. That was a mistake. I changed the "i.e." to "for example tertiary" .--PeeWee32 (talk) 08:38, 23 March 2014 (UTC)


What if a cycler has to use the cycle track in one direction, in the other he has to drive on the road ? (TheFive)

Do you know the ":forward/:backward"-tagging scheme? Example: bicycle:forward=use_sidepath for a compulsory cycleway that has only been used in the direction of the way-direction of the road. --MasiMaster (talk) 12:41, 10 May 2014 (UTC)
Yes. May be a hint / reference to this standard could be usefull. (TheFive)

Mapping for the router

Isn't it mapping for the router ?

For me the situation that there is a bicycle sign is mapped complete with bicycle=designation on that road.

If local low does not allow to use a highway near a bicycle=designated highway, than this is no data for OSM. It has to be interpreted by the Cyclist, or if he need technical support by a router, by the router itself. (TheFive)

No, the compulsory-cycleway-sign tell the access rights. The right belongs to both the cycleway AND the road. bicycle=designated only can't work because it's not well defined and there is a missing relation to the road to the tag belongs. It can also be used for cycleways with any bicycle-sign or symbols at the road. It is also used in countries which have no compulsory cycleways. And yes, access-tags are also used by routers to find a valid route. Where is the problem, or should we don't tag all access-tags, because it is mapping for router? --MasiMaster (talk) 13:04, 10 May 2014 (UTC)
Let me explain my thoughts.
* First: In the talk-de discussion it is often referred to bicycle=designated is not sharp enough, and does not state, wether there is a sign or not. This is may be correct, but we should sharpen the bicycle=designated tag, to have a clear situation and not introduce another tag.
* Second: For me the tag is calculable by a computer with the information that is already tagged (after solving "first"). So it is redundant. We should avoid a redundant tagging scheme whenever it is possible.
* Third: I understand, that programming routing algorithms is much more easier with this tag. I suggested to create an overpass-api like server, that computes the necessary redundant tags. This could contain also address tags that are inherited from buildings by POIs, or max height tags, that are derived from Bridges to a relevant set of highways, to improve truck routing. The tag "overaverage number of curves" is than derived for the motorbike routing. And the OSM data is not blown up by routing tags.
TheFive (talk) 15:03, 10 May 2014 (UTC)“
To begin with, please note that the legal situation in Germany regarding cycleways is different from the situation in the Netherlands. This may explain in part the continued confusion over the need for this tag. We have three different categories of cycleway, with different access rights for different categories of bicycle/mofa/moped. (Pedestrian access rights on cycleways are also different from Germany.)
Then, directly responding to the three points:
* (1) In the Netherlands, "sharpening" the =designated tag will not help, since there is also the non-compulsory bicycle path (which is just as 'designated' for bicycles as the compulsory path).
* (2) No matter how much you improve the routing calculations, there may be situations where it appears better for a bicycle to use a road for part of a route instead of a compulsory cycle path. This is simply not allowed (for ordinary bicycles).
* (3) Tagging should make clear what access is and is not allowed; it is an issue of access rights, not of routing optimization. See further points (1) and (2).
Why not simply add a bicycle=no tag to roads (as indeed we have been doing up till now in the Netherlands)? Because there are types of bicycles which are allowed to use EITHER the compulsory cycleway OR the road (e.g. bicycles with a trailer, etc., etc.). However, there are also roads which are explicitly signed to exclude ALL types of bicycles. These two different situations need different tagging. Without the additional tag, in the Netherlands, it is not possible to give accurate information about access rights to roads with adjacent compulsory cycleways.
Tagging access is not "tagging for the router", it is just tagging what the status is of a path or road. Frankl2009 (talk) 18:28, 11 May 2014 (UTC)
I do not know the situation in the netherlands (only that all citizens are biking a lot). So i have to switch to a more abstract discussion, and may be do abstain from the voting.
If there are different kind of cycleways with different impacts on riding on the "main road", this status should be mainly be given by signs. As we should focus on "Mapping whats on the Ground", this should be mapped as accurate as needed. May be the current tagging scheme does not cover this need. From an abstract point of view - the situation is same as in germany. There is a road A with a given status. Nearby there is a road B, and because A fulfills the rule BIKE, B is inheriting the rule NOBIKE. Would A fullfill SPECIALBIKE than B can inherit NOSPECIALBIKE. If SPECIALBIKE <> BIKE you need 2 tags on B. And what is with pedestrians ? I assume they are not allowed to walk on the highway, if there is a footway nearby. If there comes up a need to mark the highway itself with foot=use_sideway, what is the third tag.
So in summary, my fear is still, that we are introducing redundant tags. "TheFive (talk)"
One problem in the road A, road B example is that it is not necessarily clear whether roads which are nearby are related and therefore trigger the 'don't cycle here' rule (this would, I guess, require a relation,making it even more complicated for mappers, especially in countries with extensive cycleways). In the absence of a relation, the navigation software would therefore have to infer from proximity whether the rule applies.
Note also that there are cases where a parallel cycleway is not compulsory even though it is very close to a road and therefore there is no restriction in such cases on cycling on the road.
I believe it is asking too much of such software to get it right all the time. And I repeat that it is not a question of optimizing routing but of PREVENTING an illegal route being offered. It is an access question like any other.
Regarding the question whether one should map only on the basis of traffic signs on the ground: the meaning of the traffic signs is given in the highway code in a country and the signs only have meaning with regard to that code. Different countries use the same sign but with slightly different regulations (e.g. the categories of vehicle to which a sign applies). For example, the sign for "no motorized vehicles" which in many countries really means none, including mopeds/mofas, does not have this meaning in the Netherlands where it does NOT apply to mopeds (and mofas). You cannot map traffic access without taking the actual meaning into account. Such rules could be held as country defaults in a suitable place and used as necessary by routing programmes, but we are nowhere near this situation yet, imho, so I am in favour of explicit tagging of access.
The Dutch rule regarding cycleways and pedestrians is this: If there is a sidewalk, you walk on the sidewalk. If there is no sidewalk, but there is a cycleway (of any of the three types), you walk there. If there is no sidewalk and no cycleway, you walk on the (side of) the road. You are not allowed to walk where there are appropriate signs forbidding pedestrians (specific cases) or on motorways. You are right that if you want to make this explicit for routing purposes, it would need tagging. The work done in Germany especially on tagging for the blind, etc., indicates to me that tagging needs to be very specific indeed. However, that is outside the scope of the present proposal and I think it does not help to include that discussion here. Frankl2009 (talk) 11:44, 19 May 2014 (UTC)