[OSM-talk] Instead of voting

Matt Amos zerebubuth at gmail.com
Fri Oct 9 11:49:27 BST 2009


On Fri, Oct 9, 2009 at 5:06 AM, Russ Nelson <nelson at crynwr.com> wrote:
> Doctau created the following page, and various other people have
> contributed to it.
>
> http://wiki.openstreetmap.org/wiki/Proposed_features/VotingOnTheWikiIsStupid
>
> I don't think voting is stupid, but I do believe that voting is not
> productive.  Here's what I believe we should do instead of voting on
> features:
>
> 1) Just map.
> 2) Use existing keys if you can.
> 3) Use existing tags if you can.
> 4) If you used a tag that isn't in the wiki, document your use of the
> tag, so that other people won't use your tag to mean something else.

this is awesome advice. if i could add to (4) that any new tag ought
to be verifiable (http://wiki.openstreetmap.org/wiki/Verifiability),
as this helps avoid confusion and edit wars in the long-run.

there seem to be several different aspects to tagging that the current
tag voting procedure seems to conflate; use, documentation and map
features. separating these out, i think:

1) tags don't need to be voted on in order to be used. this is just
common sense - there's nothing in the editors to prevent free-form
tagging and i'm sure we don't want to stop people free-form tagging.
that's part of what makes OSM genius and unique.

2) tags shouldn't need to be voted on in order to be documented. i
don't see why we would ever want to prevent anyone documenting
anything. documentation is good, right? especially if it comes with
pictures.

3) the inclusion (or not) of a tag on map features may well be
something that it is worth voting on. it could be done on a purely
mechanical basis by counting the tag usage in the database, but this
is somewhat lacking in reason and flexibility. harry wood suggested
some useful ideas in his SotM talk
http://www.harrywood.co.uk/blog/2009/10/04/community-smoothness/ .
certainly, though, we should assume that not all tags make it onto map
features, not even most of them, but a small set of the most commonly
used / most important (fsvo "important").

> 5) If you disagree with the definition of the key or value, then
> create a new key or value with a different name, use it in your
> editing, document it in the wiki, AND (this is important) put a link
> to it in the definition that you disagree with.

from our useful chat the other day on IRC maybe we can put a set of
guidelines out there to help people resolve these competing tagging
schemes. in general, prefer the tagging scheme which:

1) preserves more information
2) is verifiable, or more easily verifiable
3) has been recommended by respected members of the community

for (3) we're back to harry's talk about how do we, as a community,
recognise those respected members?

> 6) The risk of this system is that people will not find tags that have
> the meaning they're looking for.  They'll then create a new tag which
> has an identical or similar meaning to an existing one.  If you find a
> pair of these tags which have similar meanings, you should edit the
> wiki pages for them, and include pointers to each other.

and possibly a link to tagwatch/osmdoc/tagstat so that people can find
out which is more often used in practice.

> The benefit is that people spend more time mapping and less time
> coordinating with each other on things that don't need to be
> coordinated in advance.

+1.

cheers,

matt




More information about the talk mailing list