[Tagging] Planning route in the shade during hikes either in urban areas or forests

[bkil]

> I suppose the MVP is actually to assume that if a path goes through a wood then it's shaded, and if it's within a distance north (or south) of some trees, then it's shaded.
> To be more specific than this, shade
> would anyway need custom analysis or rendering, so it seems like implementing this heuristic would be an easy first step.
>

That's a good observation, but this is so easy that humans regularly
are doing this with there eyeballs anyway (I am when I am organizing
hiking tours). What we don't have right now is tools to help this (and
data to help these tools). Also, I think most humans spend 95% of
their times within their settlement, not in the woods, so implementing
this heuristic will help, but only a small minority. I also regularly
map vegetation, parks and trees in villages, and I must say the map
around here is still blank in this aspect. For all you know, there are
roads and a few houses here and there and some POI, but no greet
whatsoever (according to the map at least).

> In cities, the same would apply for buildings of a certain size.
>

Surely this depends where you live, but many live in rural areas and
smaller towns where houses with gardens are the norm and many streets
have all kinds of vegetation planted nearby some, but not all of the
sidewalk (village green). Many town planners call for low rise
buildings, and that they be places further away from sidewalks, hence
I would say that shading by foliage has a much more substantial
coverage than shading by buildings around here.

> If an area is seen to be shaded at a particular time and it isn't because the wood/tree/building is missing or not the right height/geometry, then that problem can be fixed. It's not about mapping all objects precisely, just mapping well enough that that shade is captured (e.g. levels of buildings).
> That's already possible with the current map. F4 map also does shade.
>

Not sure whether you followed the whole discussion, but this was
already answered. Yes, mathematically and theoretically you are
absolutely right, but the real world begs to differ. It won't work in
areas where you did not draw buildings, their shape and their height
(99% of the Earth or so?)

The question was which one is easier and more realistic to have it
finished in a campaign, drawing (and maintaining) all human structures
and all vegetation, or adding such an approximation? The latter is
orders of magnitude cheaper. The former is also on open project and
I'm not against that we finish that, but we have been struggling with
that for decades now and it doesn't look like it if we will be
finishing in the coming few decades with it. But we could start
collecting summer_midday_shade=* where it matters right now - we
wouldn't have to wait for anything or anyone at all, and the lack of
dependencies is the best friend of a project manager.

> Shade is also visible in hires imagery, so if the date and time of the imagery is known I could imagine eventually setting up a street complete task to prompt for missing/wrong height objects because the OSM objects don't shade the right area.
>

Yes, that occurred to me as well and I even proposed it some years ago
as a research project, but didn't get any supporter. Would be a lot of
work in many ways, though.