[Tagging] Solving the tag chaos (was: Re: Government buildings)

Steve Bennett stevagewp at gmail.com
Tue Dec 8 01:13:30 GMT 2009

On Tue, Dec 8, 2009 at 9:01 AM, Roy Wallace <waldo000000 at gmail.com> wrote:

> All of these practices help to push OSM taggers towards consistency.
> *In the absence of a centralised authority*,

Now, is that absence of a centralised authority by accident or by design?
I'm obviously proposing just such a thing. Why would you not want that?

> these practices are
> necessary and, I think, effective. They complement each other. It's a
> dynamic process, as it must be.

Yes, but there are all kinds of dynamic processes. It doesn't have to be
this chaotic, or involve so much feedback in different directions.

> What do you mean? I think the only problem is when there is *more than
> one meaning per tag*. This is a problem and I'd be interested if you
> have any ideas on how to avoid it.

Here's how to cause it:
1) Have multiple authorities
2) Encourage people to use any tags they feel like, and to build communities
around such use
3) Encourage renderers to support tags that people use

I don't know of any specific examples, but it seems virtually unavoidable in
this model. Some people will use tag X to mean one thing (and renderer A
will support it). Meanwhile some people will use tag X to mean another (and
renderer Y will support it). Generally this is probably happening in
different countries, so no one really notices until quite late, and then the

> If you remove the word "randomly" from that sentence, I strongly
> disagree. Tag the world - verifiably and explicitly - and the users
> (renderers/routers/etc.) will follow. Inventing tags as necessary to
> mark a feature with no documented tag is fun, and very useful. Just be
> verifiable, and explicit. Use note=* to explain if necessary. The main
> thing is to avoid the need for others to re-survey.
Is it useful though? I'm looking through tagwatch and seeing lots and lots
of tags that are not, and probably never will be, supported:
built=*,bus_depot=yes, city_served=*, closest_town=*,complete=no,
completed=no, completion=*, continues=*,...

So, while I support the idea of an organic process for inventing new tags,
there need to be processes for collating and cleaning up that organic data,
guiding it to the centralised authority, and feeding it into the software.

Anyway, I still need to learn more about how those processes are currently
happening. I'm thinking that published lists of what tags each renderer
supports would be very handy in a couple of places.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tagging/attachments/20091208/7b45f396/attachment.html>

More information about the Tagging mailing list