[OSM-dev] GSoC: formalising knowledge in OSM

Erik Johansson emj at kth.se
Sat Apr 4 13:03:15 BST 2009

2009/3/25 Andy Allan <gravitystorm at gmail.com>:
> 2009/3/25 Radomír Černoch <radomir.cernoch at gmail.com>:
>> Firstly create the initial ontology to support at
>> least all official map elements.
> Hi Radomír,
> There's no such thing. In fact, the lack of a fixed/official ontology
> is held as one of the key defining attributes of the success,
> flexibility and scalability of OSM, so I'm not sure how well your
> project will work in the long run.

This has been repeated so many times, but in reality large parts of
the ontology has been "official" for quite a while now, haven't it?

The largest problem lies in tags that aren't used much. The meaning of
tags are not just decided by "key=value", but also who tagged it and
when the user tagged it. Since every user will interpret tags
differently, so you can never be sure exactly what was meant. E.g.
What exactly is "landuse=allotments"? Can you walk on

I think Richards idea is one of the better ones if you want to
continue the anarchistic tradition of OSM, every user has their view:

More information about the dev mailing list