[Tagging] The endless debate about "landcover" as a top-level tag

Michael Andersen osm at hjart.dk
Thu Jun 7 18:13:24 UTC 2018

For many years now I've been pretty happy to use landuse=forest pretty much 
everywhere I found a group of trees. Yes, in some cases the semantics irked me 
a bit, but landuse=forest always rendered fine. I used what worked for me.

On many occasions however I've seen newbies remove or retag landuse=forest 
areas as the very first thing they do after registering. "It's not a forest" 
(whatever that means) they argue, even if the area in question is completely 
covered with trees and sometimes even has "forest" as part of the name. On 
some occasions it's been a real hassle trying to explain that landuse=forest 
basically just means that the area is covered by trees, no more, no less, and 
that we use it because this is what renders, not because the semantics are 

So what I'm trying to illustrate is that while I'm happy to use landuse=forest 
myself, I do see a practical problem with it; Newbies taking it a bit too 
literally. If landcover=trees would render, I imagine it would make my job a 
lot easier.

torsdag den 7. juni 2018 18.41.26 CEST skrev Mateusz Konieczny:
> I responded to "what's wrong with getting rid of these bad choices?"
> I see nothing wrong with using landcover=trees.
> But in context of this discussion I understood "bad choices" as
> landuse=forest. BTW, I am happy to fix some broken tagging - for example I
> am regularly hunting downdemolished=yes (for example on buildings),
> made/make multiple mechanical editscleaning up tags (yes, it is approved by
> local community and follows mechanical edit policy).
> But deprecating landuse=forest of redefining lanes tag is not going to
> happen.
> 7. Jun 2018 16:11 by pelderson at gmail.com <mailto:pelderson at gmail.com>:
> > Rendering landcover=trees is not the same as deprecating landuse=forest.
> > It just offers the option to tag tree-covered areas on a different landuse
> > such as industrial, military, residential or commercial. I do expect a
> > shift from landuse=forest to landcover=trees, as soon as it would be
> > rendered. Not because of retagging of all forests, but because of tagging
> > the smaller treecovered patches mainly in residential areas, which are
> > now either mistagged as forests, orchards, parks and gardens, or are not
> > tagged at all because it isn't landuse as it is defined in the wiki.> 
> > 2018-06-07 15:36 GMT+02:00 Mateusz Konieczny <> matkoniecz at tutanota.com 
<mailto:matkoniecz at tutanota.com>> >:
> >> 7. Jun 2018 11:53 by >> selfishseahorse at gmail.com 
<mailto:selfishseahorse at gmail.com>>> :
> >>> On 7 June 2018 at 10:46, Christoph Hormann <>>> osm at imagico.de 
<mailto:osm at imagico.de>>>> > wrote:
> >>>> There are tons of established tags in OSM where the key makes no sense
> >>>> at all.  Don't get me started on 'waterway' for example.  But that is
> >>>> how OSM works.  Get over it, accept that people have made bad choices
> >>>> of keys when choosing tags and concentrate on encouraging and helping
> >>>> people to choose suitable keys when newly creating tags (in a
> >>>> productive way of course, not just by rejecting any idea as bad).
> >>> 
> >>> And what's wrong with getting rid of these bad choices?
> >> 
> >>  Cost, effort and confusion is not worth positive effects.
> >> 
> >> 
> >> 
> >> 
> >> Revolutions are really rarely worth costs.
> >> 
> >> 
> >> 
> >> 
> >> Making tagging more consistent is not one of this cases.
> >> 
> >> 
> >> 
> >> 
> >> Improvements are possible but not when it starts from "deprecate
> >> landuse=forest because it is not used to tag land use".
> >> 
> >> _______________________________________________
> >> Tagging mailing list
> >> Tagging at openstreetmap.org <mailto:Tagging at openstreetmap.org>
> >> https://lists.openstreetmap.org/listinfo/tagging
> >> <https://lists.openstreetmap.org/listinfo/tagging>

More information about the Tagging mailing list