[Tagging] RFC - Feature Proposal - area of steps for pedestrians.

[Tobias Knerr]

On 03.05.19 18:20, Paul Allen wrote:
> Or this https://goo.gl/maps/TxVMau8EBrLUAaeU6
[...]
> You will note that there are narrow, normal-size steps within big
> steps.  It's complicated.

Yeah, there's absolutely no chance to make that guess from just a
polygon OR the relation originally proposed in this thread. :)

On 03.05.19 17:49, Martin Koppenhoefer wrote:
> Can it cope with cases like these?
>
http://1.bp.blogspot.com/--QaPv-T1b2Q/TWEia7iPWYI/AAAAAAAAEwQ/dtB1TfiqeA0/s1600/Dresden_Treppe_BrTerrasse_1905.jpg

Only if we split it into a bottom and top part:
http://tobias-knerr.de/upload/Step%20Polygon%203D%20Examples/br%C3%BChl.png

As splitting is likely necessary to properly model the landing in the
first place, you may or may not consider that acceptable.

> Or similarly these
> https://t-ec.bstatic.com/images/hotel/max1024x768/504/50497971.jpg

Also would have to be split into multiple parts (three, to be exact) for
the algorithm to work, unfortunately. The multiple quasi-right angles
make this shape ambiguous.

> This one shows a common problem, occurring when the surroundings are not
> completely leveled (steps that do not run across the whole width):
>
http://www.bilderbuch-koeln.de/bilder/k%C3%B6ln_altstadt_nord_freitreppe_am_dom_treppe_domplatte_bahnhofsvorplatz_753a433380_978x1304xin.jpeg
[...]
> https://c2.staticflickr.com/4/3902/14798079901_ff9a5b72c8_b.jpg

The second example will work fine, the first one won't. This, too, is a
limitation shared by the relation originally proposed in this thread, by
the way.

Tobias