-
Notifications
You must be signed in to change notification settings - Fork 819
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add rendering for cycleway areas #3342
Comments
I’m against this, it doesn’t make sense semantically, likely a tagging error where area:highway=* intended
|
I agree with @dieterdreist , it should be area:highway=*, as is also used in the Polish renderer you reference via the link. |
We render other types of That's inconsistency either way. I prefer rendering only |
sent from a phone
On 14. Aug 2018, at 02:10, kocio-pl ***@***.***> wrote:
That's inconsistency either way. I prefer rendering only area:highway=* (which means dropping highway=* + area=yes for all of these cases) as a proper solution.
you seem to assume that area:highway=* had the same meaning as highway=* and area=yes
|
Yes. And what do you think about it? |
area=yes on a highway defines a highway area, i.e. an omnidirectional 2 dimensional space (free flow) as opposed to a directed, linear road. area:highway is about the surface area of all kinds of highways (in the former case it would be redundant)
|
@dieterdreist @mboeringa @kocio-pl See: https://lists.openstreetmap.org/pipermail/talk/2018-August/081102.html As we render footway/ pedestrian/ path areas, cycleway is an illlogical "gap" here, but like @kocio-pl said, we should avoid duplicating tagging schemes, so we should choose the proper one and drop the second. Dropping of combinations and render a:h=* is also good solution for me, I just want us to establish one propper tagging scheme supported by default style. |
sent from a phone
On 14. Aug 2018, at 07:49, Tomasz Wójcik ***@***.***> wrote:
I started a discussion about area:highway=* for footways etc. on TL some time ago, and the output was that using combination of highway=* + area=yes is not a tagging error.
See: https://lists.openstreetmap.org/pipermail/talk/2018-August/081103.html
Adding area=yes tag for objects which could also be linear (eg. barrier=hedge) is a common agreed method of tagging them.
As we render footway/ pedestrian/ path areas, cycleway is an illlogical "gap" here.
people don’t use it. There are less than 1000 (i.e. it is not listed as a combination) highway=cycleway with area=yes but more than 7700 area:highway=cycleway according to taginfo.
I agree hedges may be different, still what I wrote about the different meaning is established interpretation at least in Italy.
If your reading of the tags as synonyms would become widespread, it meant that it would become unclear whether it is an area where you can walk/drive as you like, or a linear street where you have to keep within your lane.
|
I don't think this is just in Italy, this is the more general interpretation in many countries, including I think in Poland based on the tagging applied there. A highway=pedestrian + area=yes can in fact be subdived in many small areas of area:highway=pedestrain + surface=x. E.g. imagine a big plaza tagged with highway=pedestrian + area=yes + name="OUR_PLAZA" to mark out the entire extent of the plaza. This plaza could have several different areas with a different surface. E.g.:
As you can see, in most cases, it is highly advisable to add the surface=x tag when using area:highway. Also note that each feature is distinct here. The highway=pedestrian + area=yes + name="OUR_PLAZA" feature is not equivalent to any of the area:highway=x features. |
@mboeringa @kocio-pl Can you join to ML discussion from link above and present your arguments for area:highway=* scheme? There is not so much people strictly decided on combination scheme, so I think we can make it for a:h=*. |
@Tomasz-W - do you have a real-world example which is tagged cycleway + area=yes that is dedicated to omnidirectional cycle traffic? I haven't seen such a case so far. People on foot have a slower mode of locomotion, thus an omnidirectional area makes sense: as a pedestrian area to stroll from shop to shop, or as a widened areal in the middle of a park. On a service way, area=yes is often used in industrial estates where you have large areas for trucks to manoeuvre. I mean, we don't need to code a case that is just a theoretical combination of keys. |
@polarbearing You seem to don't understand the idea. Way's direction, its access, wheelchair access, lit or no-lit, etc. is tagged in highway=* way (which correctly mapped is "flowing" through area middle), and area:highay=* area is just for visual area shape representation, so way's direction doesn't matter to area tagging. |
I'm not ready to discuss it outside this ticket yet. The arguments given here are interesting and I have to think about them more:
On the other hand if h=c + a=y is not wrong, I want to have it included to cover complete highway solution. I don't like adding rendering case by case without thinking about the whole system, I think this is old bad habit. |
Correct, but your original issue was to render area=yes.
No it just indicates that this feature does not exist in the real world for cycleways, otherwise please point to an example of such feature.
Yes. the first is the shapes regardless of traffic, the second describes omnidirectional traffic. |
@matkoniecz You have been quite active on this topic (OSM edits converting area=yes to a:h=*, forum posts), so I'm tagging you as you may be interested in ;) |
I don't think so looking at numbers: https://taginfo.openstreetmap.org/tags/area=yes#combinations 30.55% | highway=* It looks like ~21% (2/3) is about omnidirectional traffic, but I guess residential or unclassified are not. So this tag seems to be used for different uses and I guess rendering some of them makes the bias toward omnidirectional traffic, but it's still just 2:1. |
Trying to understand what you mean. Taginfo shows that area=yes is used in a lot of unnecessary combinations, it makes no sense e.g. on natural=bare_rock as that is an area by definition. This might come from bad imports. area=yes on a highway is supposed to describe the omnidirectional traffic. thus I think we can close the issue. |
sent from a phone
On 24. Aug 2018, at 23:40, polarbearing ***@***.***> wrote:
thus I think we can close the issue.
+1
|
I think that rename the issue to "Add rendering for area:highway=cycleway" would be better. |
That's a completely different thing and a dup of #180. |
That ticket was about adding rendering for whole area:highway=* key, and this one is for area:highway=cycleway only, so it's not a duplicate. |
Well but I'd not start rendering area:highway=cycleway as an exception without deciding about area:highway=* as a whole. |
area:highway=* is a micromapping and is totally an other strategy noboby area:highway have a representation a moment. There is an other problem with overlapped same data one line and one closed line with area = yes for exemple : |
The combination |
Combination of
highway=cycleway
+area=yes
is currently not rendered at all. As we render areas of:I think cycleway areas should be included too. At least with the same filling as the ones from list above. Another options are to render them with some red filling like in this visualisation:
http://osmapa.pl/w/area/?lat=51.09433&lon=17.0173&zoom=19&ol=Qq
or with a little bit darker gray shade than footway areas.
Edit: add rendering for area:highway=cycleway is also good solution for me, I just want a complete rendering of one of these tagging schemes.
The text was updated successfully, but these errors were encountered: