[OSM-talk] Scientific/Species Data in OSM Database - Collaborators

Martin Koppenhoefer dieterdreist at gmail.com
Mon Oct 22 23:57:40 BST 2012


2012/10/22 Peter Wendorff <wendorff at uni-paderborn.de>:
> I guess, that most species don't have fixed areas, where one can say
> "species X resides inside this area, but not outside".


for flora you could tag e.g. rare plants, mostly they don't move away ;-)


> I think, it's more like a "this is the core area for that species, and
> there's a fuzzy margin around where they sometimes occur.".
> Features like that have discussed before already, e.g. for fuzzy
> geographical regions like the alps etc., but there's no good solution yet,
> as far as I know.


I like the idea of a relation type which has other objects (nodes,
maybe also ways) included with the roles inside or outside (optional).
You won't have to put many objects in. The clue is that the extent of
the area is calculated later based on these features, so you won't
have much load on the db (especially if you keep the amount of members
small). An algorithm would find the rough area where all (or almost
all) inside-role-objects are inside and the others outside. You won't
have to draw additional elements (outline) around the area, and you
won't have to have a closed polygon, so it might be much more stable
than for instance multipolygons or routes. You also won't have a
cm-precision which you don't need. You would simply get an idea how
big the area was, and where it is more or less.


cheers,
Martin



More information about the talk mailing list