cycleway with a foot=designated and footway with a bicycle=designated distinction seems useless

[VladimirMorozov]

Cycleways can have a foot=designated tag, but still they will be rendered with a darker blue than a footway with a bicycle=designated. Both of those cases are used for paths like this:

As a result this color distinction does not help at all, only says how the way was initially mapped.

https://www.cyclosm.org/#map=18/59.43344/24.81383/cyclosm

I'd propose to use darker blue only for a bicycle-only paths with a foot=no or empty. And lighter blue for combined paths, both for a cycleway with foot=designated and a footway with a bicycle=designated.

I understand that countries with good bicycle infrastructure have many bicycle only roads and don't have many roads with combined use and maybe this rendering problem is not noticed, but this is not the case in my country. Most of the roads which cyclists are expected to use are also footpaths. And some map them as a cycleway, others as a footpath resulting in color difference which gives no information. Or maybe there is a case in some country where current rendering does have a meaningful information?

[quasart]

As discussed in #397 I agree with you, but @Florimondable has disagreed in #397 (comment)

Since this disagreement was lacking of explanations, and the topic is coming back, let’s discuss this point again. I’m still convinced it would be a more coherent view of osm data.

[VladimirMorozov]

Oh, thanks, didn't notice this issue, since the title was about something different.

I too think that it's more important to document legally verifiable truth than what some mappers in France feel. I don't even cycle in France. And it's better to support all tagging schemes users everywhere actually use instead of trying to encourage to map for the render.

Footways, cycleways, tracks should be rendered similarly, by checking who they are designated for, with footway and cycleway having defaults.

Edit:
I should also add, that if we want to render segregation, then highway type shouldn't play a role. Segregation can be added to footways too. And I still think it's more important to show which paths are free from pedestrians. In my country pedestrians and cyclists don't care about segregation much - often groups of people of more than 2 cannot fit on their side even if they want to. I understand that this may be useful on wider roads and in countries with stricter cyclist-pedestrian culture, but then I'd say it's better to have a scheme with 3 colors: bicycles only > segregated way > non-segregated way.

Edit2:
Got a bit confused at the end, we already have a variant for segregated cycleway. But I think we should treat it the same as footway with bicycle=designated and segregated=yes.

Also, I'm not quite sure with what exactly @Florimondable disagreed with.

[Florimondable]

If you have a cycleway why we should taking care of having pedestrian allowed or not ? This is a cycle map.

If there is lot of pedestrian on a cycleway, may be it shouldn’t be tagged as a cycleway but as a shared path.

Real world examples are welcome.

[quasart]

If there is lot of pedestrian on a cycleway, may be it shouldn’t be tagged as a cycleway but as a shared path.

This may be your personal preference but it’s not something we may take for granted. Did you read https://wiki.openstreetmap.org/wiki/Path_controversy ?

[VladimirMorozov]

@Florimondable because highway=cycleway with foot=designated is a perfectly valid tagging. The real world example is the road sign I posted in the first comment and it can be tagged differently. And all valid tagging schemes should taken into account.

There is no way a mapper can easily and objectively measure how many pedestrians use the shared path, especially considering that it changes with time. Cycleway and footway tags don't say anything about how many cyclists/pedestrians use the path.

What you want the highway=cycleway tag to be does not align with what it is.

Since it is a cycle map I think cyclists would benefit from knowing if pedestrians are to be expected on a highway=cycleway.

[Phyks]

highway=cycleway with foot=designated is a perfectly valid tagging. The real world example is the road sign I posted in the first comment and it can be tagged differently. And all valid tagging schemes should taken into account.

If I'm not mistaking, this is case S7 in https://wiki.openstreetmap.org/wiki/Bicycle which indeed is a valid tagging scheme worldwide.

Note that this is however different than the "segregated" tagging scheme which is S5 in https://wiki.openstreetmap.org/wiki/Bicycle and explicitly direct to use highway=path for such cases.

EDIT:

Since it is a cycle map I think cyclists would benefit from knowing if pedestrians are to be expected on a highway=cycleway.

Sure, but I'm not sure we should be trying to guess what the infrastructure actually is, beyond standardized tagging schemes:
1/ First, it starts to become an issue at the scale at which CyclOSM aims to operate (serve "live" tiles at a global scale), where performance becomes an issue. The more branches, the more performance becomes an issue.
2/ Second, it's not helping moving towards the standardized tagging scheme which is easily consumed by everyone using OSM.

So, TLDR, of course, I am in favor of rendering everything as it is expected from the tagging description, as long as the scheme is somehow standardized (or can be considered as such due to a wide use locally).

[Phyks]

By the way, could we try to limit the number of similar issues. The discussion here seems to cross the one in #397 on a number of subjects. Could we try to sum up the main concerns here and if #397 and this one are indeed about the same concern, merge the discussion in a single thread? Thanks!

EDIT: See as well #450.

[Florimondable]

@Florimondable because highway=cycleway with foot=designated is a perfectly valid tagging.

yes

The real world example is the road sign I posted in the first comment and it can be tagged differently. And all valid tagging schemes should taken into account.

That’s not a real world example, it’s just a board, anyhow this board doesn’t describe a cycleway but a path.
There is no "valid" schemes in OSM, ;)

There is no way a mapper can easily and objectively measure how many pedestrians use the shared path, especially considering that it changes with time.

There are ways to measure that.

Cycleway and footway tags don't say anything about how many cyclists/pedestrians use the path.

What you want the highway=cycleway tag to be does not align with what it is.

Since it is a cycle map I think cyclists would benefit from knowing if pedestrians are to be expected on a highway=cycleway.

Pedestrian is always being expected on cycleway, the question is how many.

If I remembre well the choice of not taking care of foot=designated on cycleway is that if it’s a cycleway it’s cycleway state is more important than knowing that maybe there are pedestrian on it.
And the choice that if a path is shared it should be tagged with path not cycleway value.

[VladimirMorozov]

If I'm not mistaking, this is case S7 in https://wiki.openstreetmap.org/wiki/Bicycle which indeed is a valid tagging scheme worldwide.

Note that this is however different than the "segregated" tagging scheme which is S5 in https://wiki.openstreetmap.org/wiki/Bicycle and explicitly direct to use highway=path for such cases.

About segregated: in addition we have this https://wiki.openstreetmap.org/wiki/Tag:highway%3Dcycleway where it doesn't prohibit tagging foot access and segregation. Also, in ID editor template for the combined foot and bicycle access is based on a cycleway. So not only we have already a lot of combined ways mapped as cycleway + foot=designated, this is encouraged by the editor and will continue to be added. Same goes for StreetComplete, if it sees a cycleway with foot access, then it asks if it's segregated and doesn't care about changing it into a path or a footway.

So I'd say we have to live with that the data in OSM is and will be like this for a while.

First, it starts to become an issue at the scale at which CyclOSM aims to operate (serve "live" tiles at a global scale), where performance becomes an issue. The more branches, the more performance becomes an issue.

There is a branch currently and I'm arguing that it is useless in a lot of places and depends on a personal opinion of someone on what is a cycleway. The idea is to change this branch's condition. I haven't looked at the query, but are you sure this will be a big performance hit compared to all the other things which are done to render something? It's just that I think it is a good feature for a map which focuses on cycling. Currently for me the map is colored in random blue shades and all I know is that somewhere in France they may mean something. Think of the end user who doesn't know about tags at all, he'd be confused why some combined ways are lighter and some are darker.

That’s not a real world example, it’s just a board, anyhow this board doesn’t describe a cycleway but a path.
There is no "valid" schemes in OSM, ;)

it's not helping moving towards the standardized tagging scheme which is easily consumed by everyone using OSM.

There are no valid schemes, but there is an actual state of how the data is and how it will be. I don't think there is a point of going against something this big, you won't win. You think this is a path, someone thinks this is a cycleway with foot=designated, then someone else thinks it's a footway with bicycle=designated. Tagging could be better, obviously, as mentioned in "path controversy", but there are no changes on the horizon and if there were, they would be backwards-compatible. The point is, all those different tagging schemes document the same thing and it should be drawn the same.

As such this

this board doesn’t describe a cycleway but a path.

is a personal opinion. I'd like it to be like that, but it's not and it's not really encouraged by the wiki.

Pedestrian is always being expected on cycleway, the question is how many.

In Estonia it's forbidden to walk on a cycleway and while some do walk, there is usually a footway in parallel where most people walk. Don't know about France, but I've heard they are very strict with rules like this in Netherlands.

There are ways to measure that.

Yes, but they require a lot of organization, maybe from government, or a lot of time from mappers. This is not something which can be achieved in short time and to be expected by default for every way. It would definitely also require separate tags (not like there would be some threshold by which we decide if it's a cycleway or a path). If we could have this data for most places, this would be awesome and I'd say it would be more important than seeing if the cycleway is combined with a pedestrian path. But unfortunately, this is not possible in near future for most of paths.

[Florimondable]

this board doesn’t describe a cycleway but a path.

is a personal opinion. I'd like it to be like that, but it's not and it's not really encouraged by the wiki.

Real world example here : https://www.google.fr/maps/@59.450609,24.8606561,3a,75y,2.13h,79.89t/data=!3m6!1e1!3m4!1sYwzbchIZ_RZ3_F5U4AKYNA!2e0!7i13312!8i6656
https://www.cyclosm.org/#map=17/59.44981/24.86211/cyclosm

I don’t understand how this can be map as a highway=cycleway, it’s a shared path by design and defined by the board. So for me this is just mistagging and could be fixed.

[quasart]

This is exactly what we say, « for you » it’s wrong. JOSM tags your example as a path, you’re not alone :-)

But many mappers think highway=path is a bad tag, and it should be avoided when possible. ID editor tags your example as a cycleway.

Both tagging styles coexist, we don’t understand why you assume second one is bad tagging.

It’s not even a French preference (may be Paris only?) I can find many « voies vertes » tagged as cycleway if you want.

If for some reason I missed that highway=cycleway foot=designated means something important somewhere, maybe we could add some kind of dark blue outline to make tagging visible ? But AFAIK both mean the same thing

[VladimirMorozov]

Here are some examples that this tagging of combined ways as a cycleway is actually used (the green ones are cycleways with foot=designated):

Copenhagen

Stockholm

Edinburgh

[out:json][timeout:25];
(
  way["highway"="cycleway"]["foot"="designated"]({{bbox}});
  way["highway"="footway"]["bicycle"="designated"]({{bbox}});
  // way["highway"="path"]["bicycle"="designated"]["foot"="designated"]({{bbox}});
);
out body;

>;
out skel qt;

{{style:

way[highway=cycleway]
{ color:green; fill-color:green }

way[highway=footway]
{ color:red; fill-color:red  }

way[highway=path]
{ color:blue }

}}

[Florimondable]

Let's check some:

https://www.openstreetmap.org/way/25674359
https://www.google.fr/maps/@55.7124473,12.4865141,3a,75y,21.88h,73.76t/data=!3m6!1e1!3m4!1s0hmHvQdXqf8MRCpVxPKvGA!2e0!7i13312!8i6656
Bad tagging (I cycled there, there is a footway and a cycleway next to each other)

https://www.openstreetmap.org/way/105932499
https://www.google.fr/maps/@55.629581,12.4373676,3a,75y,324.83h,62.64t/data=!3m7!1e1!3m5!1s0Z9yX_QI-GmgVS2ieS2EZw!2e0!6shttps:%2F%2Fstreetviewpixels-pa.googleapis.com%2Fv1%2Fthumbnail%3Fpanoid%3D0Z9yX_QI-GmgVS2ieS2EZw%26cb_client%3Dmaps_sv.tactile.gps%26w%3D203%26h%3D100%26yaw%3D196.47913%26pitch%3D0%26thumbfov%3D100!7i13312!8i6656
Bad tagging (That's not a cycle way, it's just a path).

https://www.google.fr/maps/@59.3136572,18.0572908,3a,15.1y,110.32h,86.35t/data=!3m6!1e1!3m4!1sQDRPsZHoigdOrLTpRme2CA!2e0!7i16384!8i8192
https://www.openstreetmap.org/way/361740026
Shared path, not a cycle way.

https://www.google.fr/maps/@55.9627294,-3.196277,3a,15y,323.37h,83.37t/data=!3m6!1e1!3m4!1s47w8BQEDwAETyidAi3nh4A!2e0!7i16384!8i8192
https://www.openstreetmap.org/way/4871786#map=19/55.96297/-3.19662
Also a shared path.

Of course it's my point of view, which is very important since I'm a coauthor of this map :)
I do think tagging shared path with highway=cycleway + foot=designated has flaws, and I think CyclOSM shouldn't promote this.

[VladimirMorozov]

Ummm you don't really prove anything by those examples. I don't know what to say. Some are bad tagging, but not all. Others are bad tagging in your personal opinion.
If you think that your view is the most important and you don't care about data in OSM then so be it. I just brought arguments, I don't make you do anything.

[quasart]

Are you saying that:

• you agree this tagging scheme is used widespread
• but you don’t like it
• you will keep rendering pathes in wrong color, waiting for half of the world changing its tagging scheme

Or is there a misunderstanding ?

[Florimondable]

Many people may say that 2 and 2 make 5, that's not a reason I'll accept it.

highway=cycleway is for bicycle only way not for shared path.

You're free to fix tags of the shared paths in your area.

[VladimirMorozov]

Tagging for the renderer is considered bad. Even if it is a change from one valid scheme to another valid scheme, I don't think that it should be done.

[Phyks]

I'm jumping in into this thread.

First of all, regarding the performance argument

I haven't looked at the query, but are you sure this will be a big performance hit compared to all the other things which are done to render something?

I was just saying that CyclOSM now operates on a quite large scale, with comparatively quite small servers. I did not investigate the performance issue here, and in any case performance considerations should not be an obstacle to proper rendering according to OSM guidelines, it should only be taken into account when taking extra steps to further guess / improve tagging when it is lacking details.

Regarding the overall issue and the specific point of highway=path vs highway=cycleway, I dug in the mentionned wiki pages. @VladimirMorozov made a strong point by mentionning usage of the alternative tagging scheme #575 (comment). There is another question for me beyond usage about whether this is actual proper tagging or not. Usage might contradict wiki and tagging (hence, adding render for it in CyclOSM might promote bad tagging practice, which we certainly do not want, or at least this risk should be properly taken into account and benchmarked). When both usage and tagging are in phase and widely used / not contested globally, it's not our OSM data consumer's (as developing a map render) place to challenge the tagging IMO. So, let's review the wiki:

• Case S5 in https://wiki.openstreetmap.org/wiki/Bicycle explicitly direct to ues highway=path for segregated cycle / foot ways.
• Case S7 in https://wiki.openstreetmap.org/wiki/Bicycle explicitly allow whatever tagging highway=path|footway|cycleway for non-segregated cycle / foot ways.
• This is further emphasized by https://wiki.openstreetmap.org/wiki/Tag:highway%3Dcycleway#Foot_access_on_cycleways which definitely allow such shared ways to be tagged as highway=cycleway + foot=designated, highway=footway + bicycle=designated or highway=path, both tagging being considered as equal (note the vocabulary here, "could be tagged").
• Furthermore, there are countries where default access restrictions for highway=cycleway implies access to pedestrians, see https://wiki.openstreetmap.org/wiki/OSM_tags_for_routing/Access_restrictions#Australia. For the sake of consistency for these countries, we should probably not make a distinction between these equivalent tagging schemes.

I do not see any edit wars or specific discussion in the before mentionned wiki pages concerning this highway=path|footway|cycleway point, which makes me think it is widely adopted and not contested. I found out a concern similar to the ones expressed by @Florimondable (if I'm not mistaking) in https://wiki.openstreetmap.org/wiki/Talk:Duck_tagging, dating back from 2016 and not having more traction since.

Note: Note that the situation is very much different in France (and hence, in the FR:Bicycle version of the Bicycle page, which I'm biased towards and way more familiar with - although CyclOSM being available worldwide should not reflect this tendency) where the bicycle wiki page explicitly states to use highway=path whenever pedestrians and bicycles live together: https://wiki.openstreetmap.org/wiki/FR:Bicycle#Voies_pi.C3.A9tonnes_partag.C3.A9es. Same for the french version of highway=cycleway wiki page which minimizes the highway=cycleway + foot=designated variant of highway=path bicycle=designated foot=designated https://wiki.openstreetmap.org/wiki/FR:Tag:highway%3Dcycleway.

Note 2: The German wiki (German community being one of the most active I know about when dealing with bicycle tagging) seems to have an extensive section on the matter: https://wiki.openstreetmap.org/wiki/DE:Bicycle/Meinungsstreit_Radverkehrsanlagen. I'm not a German speaker, so I'll welcome feedbacks from German speakers, but Google automated translation seems pretty good here https://translate.google.com/translate?hl=&sl=de&tl=en&u=https%3A%2F%2Fwiki.openstreetmap.org%2Fwiki%2FDE%3ABicycle%2FMeinungsstreit_Radverkehrsanlagen.

All of this being said, I am convinced highway=footway + bicycle=designated, highway=path bicycle=designated foot=designated and highway=cycleway foot=designated should be considered as synonyms (up to an eventual segregated question) in most countries around the world.

Before having a deep dive into the code, and in particular PR #450, is there any other points (beyond the previous equivalence) that should be addressed as part of this issue (#575) or its sister issue #397?

[Booligoosh]

Just adding that I support this change. The Australian tagging guidelines on the wiki say to tag shared paths as highway=cycleway + foot=designated.

Due to most mappers in my city following this guideline, with the current CyclOSM style, it is impossible to tell the difference between shared paths vs. dedicated cycleways, and plan my route to prefer the latter where possible. Using the light blue colour for highway=cycleway + foot=designated would make things much easier.