Improve rendering of segregated cycleways & footways

[euf]

This is a follow up to #7878, #8026 and similar.

Consider two ways, the first one going west and the second one going north:

1. 34.689945, 33.016965 https://osm.org/way/1265838803
2. 34.689922, 33.017037 https://osm.org/way/1201097783

Cycle & Foot Path	Cycle Path

1. The first one is rendered as Cycle & Foot Path;
2. The second one is rendered as Cycle Path.

This is incorrect.

Both should be Cycle & Foot Path as per having

• highway=cycleway
• bicycle=designated
• foot=designated

Please note that both ways were last edited on 26 March 2024.

---

iOS 17.3.1

Organic Maps version: ⁦2024.05.04-10⁩ (TestFlight beta)

OSM data as of 29 April 2024

[pastk]

Both should be Cycle & Foot Path as per having

* highway=cycleway

* bicycle=designated

* foot=designated

The second one is segregated (segregated=yes) so its rendered like a white dashed footway and cycle one over it (it'd be better to render the cycle line just on the side, but we don't have this capability yet). In the routing it also behaves like a dedicated cycleway for cyclists and dedicated footway for pedestrians.

The first one is shared (non-segregated).

So from my POV everything is as intended.
Please re-open if you think it needs more attention!

[Misalf-git]

From my understanding both (segregated=yes/no) should be called "Cycle & Foot Path".

[pastk]

Or maybe they should be called "Shared ..." and "Segregated ..."

[euf]

segregated=yes <...> rendered like a white dashed footway and cycle one over it

At the moment the footway white dashes are totally indistinguishable below the blue cycle path.

As a pedestrian I get an impression like there's no footpath: both looking at the map (solid blue line here) and inspecting the Place Page (it clearly says Cycle Path).

I'd suggest the same rendering for both cases, no matter segregated or not, because the main questions are "Can I walk here" and "Can I cycle here". In my opinion the current approach fails to answer the first question in a meaningful way.

"Am I separated from cyclists" and "Am I separated from pedestrians" are important questions too, especially for those who care about safety. They are nevertheless secondary to the questions of accessibility in general.

Would you consider the following solution as feasible?

For classification, clearly separate these cases to

1. "Foot Path" (bicycle=no)
2. "Cycle Path" (foot=no)
3. "Shared Foot & Cycle Path" (segregated=no, assuming pedestrian priority over the cycling)
4. "Cycle & Foot Path" (segregated=yes, assuming cyclists travel freely)

For rendering, 3rd and 4th should have the same look just as segregated=no has now, being white dashes over solid blue. They could be further differentiated by some variations between them (e.g. blue parts become thicker for the 4th or white dashes become longer for the 3rd).

For navigation, cyclists could have the same weights/etas for 2nd and 4th. They get some penalty at 3rd ways. Depending on local rules they are either banned from 1st ways or just get an even bigger penalty there.

I believe that my approach is rather reasonable since it prioritizes the "can I move here" question for both cyclists and pedestrians. It makes the map more informative for a wider audience. The segregated distinction is less important in my view and should render accordingly.

[pastk]

At the moment segregated cycleway features have both types highway-cycleway and highway-footway.
This makes routing treat them as best/optimal options for both cyclists and pedestrians.
The downside is that we can't render them a special way, they just render as a footway and cycleway as if they were separate features.

To customize rendering we need to introduce a new type e.g. highway-cycleway-segregated (similar to e.g. highway-footway-bicycle we have already). But then routing will treat it as a dedicated cycleway and will give penalty to pedestrians..

@vng how to make routing treat highway-cycleway-segregated in a special way? I guess we'll have to add e.g. routing::HighwayType::HighwayCyclewaySegregated?

(there is another option to represent segregated cycleways as a combination of two new types e.g. highway-cycleway-segregated and highway-footway-segregated, then the routing logic won't need to be changed, but IMO its a hacky workaround and we'll need to improve routing logic for cycling and walking sometime in the future anyway)

[Misalf-git]

I beleive cycleways on segregated=yes may still be fairly visible with quite some transparency instead of being opaque since the underlying white dashes of footways provide high contrast.

[pastk]

I beleive cycleways on segregated=yes may still be fairly visible with quite some transparency instead of being opaque since the underlying white dashes of footways provide high contrast.

Yeap reducing cycle line's opacity might be an option.

Still to do it for segregated cycleways separately from dedicated ones (w/o a footway) we need some refactoring as I've described above.