> 22 maj 2020 kl. 17:25 skrev Andy Townsend <ajt1...@gmail.com>:
> I think that there's another problem with the standard style as well - aside 
> from surface rendering it's hugely biased towards urban centres.  Looking at 
> https://www.openstreetmap.org/#map=13/53.9023/-0.8856 you can't see any paths 
> at all at that zoom level due to the "Central European Graveyard problem" - 
> compare with 
> https://map.atownsend.org.uk/maps/map/map.html#zoom=13&lat=53.9006&lon=-0.8795
>  to see what you're missing.

Slightly off-topic for the tagging list, since it's more of a rendering issue, 
but: This is something I have been scratching my head over, too. In some areas 
of Sweden, it is perfectly possible to render all footpaths even on at 50k 
scale map, but then you come across areas like this, which just clutter up 
 I would like to render only the "main throughfares", but it's hard to know 
what they are, when everything is just highway=path. I have considered just 
running some spanning-tree algorithm and hoping for the best. :-)

The 50k "official" maps I am trying to emulate used to have a rule that paths 
were shown if they were at least 500m (I think) and led to a destination which 
was also visible on the map. However, Lantmäteriet ("OS Sweden") no longer 
maintain footpaths except those that are reported by sports clubs, local 
councils, hiking clubs etc because they lack the "boots on the ground" to do 
so. (THis is obviously a niche for OSM to fill!) Unfortunately, the "leads to" 
bit is tricky to automate given OSM data.

(Graveyards are less of a problem for me. I haven't bothered, but I could tell 
my renderer not to render footpaths in them, unless the scale is 1:12500 or 
bigger. Ideally, I would like to ignore everything but paths leading from one 
gate to one on the opposite side, but I haven't come up with a way of 
identifying those, yet. Some kind of graph analysis seems to be called for. 
Again, it would be nice with some sort of hints about that the throughfares 

Tagging mailing list

Reply via email to