On 2020-01-15 14:05, Philip Barnes wrote:
> On Wednesday, 15 January 2020, Colin Smale wrote: On 2020-01-15 13:52, Lionel
> Giard wrote:
>
> Yes this is something you can do with any distance algorithm in available in
> any GIS tool. That's not something that i would ever map as it would vary
> with any geometry change of the ways between the road the point you measure,
> added to the fact that it add nothing to explicitly map it (in my opinion at
> least). It is essentially what any routing app will calculate when you ask
> "find the nearest POI" for example.
> Surely the question is not about straight-line geometric distance/time
> but about route distance along paths etc. in pedestrian mode. "Nearest"
> is not nearest in terms of geometric distance, but the POI with the
> shortest route.
It is, but surely in calculating that distance it is easier and better
to simply map the paths?
Absolutely, but actually a router that can handle routing over a polygon
might be more versatile as it would not be restricted to the lines that
people have mapped as paths. Imagine an algorithm that can take a
polygon (town square, park, whatever), detect all ingress/egress points,
use visibility rules to get round concave bits and obstacles like
fountains and statues, create a temporary routing graph, deduplicate
common segments and otherwise optimise, and use that to route people
across the polygon. In this case the "pedestrian" can be guided to the
target POI without the necessity of having every possible pedestrian
path mapped while still respecting barriers (which we do have in OSM)
such as buildings, walls, fences, ditches etc.
_______________________________________________
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging
