[Tagging] Feature Proposal - RFC - natural=bare_rock

M∡rtin Koppenhoefer dieterdreist at gmail.com
Sat Jan 29 17:28:25 GMT 2011


2011/1/29 John Smith <deltafoxtrot256 at gmail.com>:
> You are yet to show how landcover=* makes things better. All I see
> landcover=* doing is duplicating surface=* and confusing people.


It is mainly the meaning, surface refers to the surface while
landcover refers to the general coverage. I agree that sand is a good
value for surface, but at the same time there could be
landcover=trees.


> things instead of sticking your head in the proverbial sand...
> http://wiki.openstreetmap.org/w/index.php?title=Key:surface&action=history
> Specifically:
> "(cur | prev)  2010-07-20T00:30:54 RichardMann (Talk | contribs)
> (1,883 bytes) (Post tag-list discussion tidy up) (undo)"


I took a look and I find this edit highly disputable, and indeed some
of it in the actual state of the page says now the opposite ;-)

e.g. "default" for "roads". I think that someone must be able to tell
from the data if a road is paved or not without further analysis, but
with this definition you must know for every part of the world what is
considered a "road".


> If anything surface has been in use for a very long time, why can't we
> just use it?


we could. What are the other objects already tagged with surface? What
are the suggested values for surface on other objects then ways?

I neither find this in the wiki nor significantly in the data. I could
also support surface (there might be space for landcover as well).
Actually surface=sand or bare_rock makes perfectly sense.

This has also effects for the users of the data. If you import the
data into specialized database, with only surface as key, you would
have one column less. This can be either good or bad (less columns
with the same implications would be preferably, while you would have
more effort to filter what you don't need).

...
> In either case you could still tag them both as surface=sand and they
> could render without knowing anything about the other tags being used,
> which seems to be a good thing imho...


yes, for rendering you can use whatever key-name, I'd like to think
which properties are better described. For vegetation landcover seems
more appropriate while surface seems better for "material".

...
> Sure, but the primary reason a lot of people tag stuff is to have it
> show up on a map, not so they can do statistical analysis or whatever


at a certain point, you will not be able to show everything on every
map, but there will be maps that show what you want, (probably maps
made by you). Which information/aspect and the logics how it is
diplayed are up to the makers of the rulesheet.

cheers,
Martin



More information about the Tagging mailing list