[Tagging] How to overcome lack of consensus

Murry McEntire murry.mcentire at gmail.com
Wed Sep 18 17:07:53 UTC 2013


On Wed, Sep 18, 2013 at 7:49 AM, Martin Koppenhoefer <dieterdreist at gmail.com
> wrote:

> ... IMHO if there is a definition in the wiki and someone then tags
> something with this tag you have to believe that he followed that
> definition, at least until you can find a consensus to change this
> definition. The most you should do is add a hint to the wiki that there is
> a similar tag in use and link to it, but you shouldn't imply that the
> similar tag has the exact same meaning as long as you don't know it for
> sure.
>
>
After some experience with OSM maps, my first assumption on seeing a tag
used is that the tagger used the meaning most in line with their personal
experience. When tagging something, if a term comes up in the presets that
has a meaning to the tagger, they use it - skipping terms that may be more
accurate that are unfamiliar to them - and no consulting of the wiki. The
tagger may look at another already tagged feature of the same type and
reuse the tag. As a secondary measure, they scan the features page of the
wiki, until they find an agreeable term and use it. As a last effort, they
may read the wiki page and the "OSM definition". The number of taggers that
have read the wiki page for every tag before use they've ever used is
likely minuscule.

Unfortunately, personal experience is often incorrect or very localized.
For all the words in a person's vocabulary, very few definitions were
formally checked with a dictionary; most are (sometimes incorrect)
interpretations for observed usage. . I have seen meanings for tags
defended on the mailing list they were quite different from any dictionary,
wikipedia, or other formal or common reference. "I grew up with these and
this is what it means." Relying o personal experience is dicey as members
of the same immediate family can have different definitions for the same
word.

The use of localized meanings and terms results in a map not useful to
those outside the locale when visiting - surely a poor state of affairs
when one trusts OSM for local use, but switches to Google or other maps
when outside your own locale because those maps have consistent meaning
across locales. Some uniformity makes OSM that much more useful. Please do
not sneeze at some need for consensus on tagging. Imagine an OSM that had
200 terms in use for similar entities and this existing for every tag in
OSM; where traveling 50 miles meant looking up a new set of tags and
definitions to use the map. Having 2 terms for the same entity is simply a
smaller version of the problem.

This points to the importance of attempting to pick terms that have a
primary meaning on first glance that  go with what it being tagged as
opposed to a term where the intended meaning for the tag is deep in the
multiple meanings of the word. "Plot" is a good example of a word that will
mean different things to different taggers, so should be avoided.

Given the convention of using British English, consulting the Oxford
English dictionary (or Collins or other suitable British sourced
dictionary) would be the conscientious methodology. Look at synonyms for
less ambiguous terms. I would also look at American dictionaries to see if
another term avoids British/American ambiguities (not always possible).
Translation dictionaries are poor sources of definitions as they often
loose the more common meanings of words or pick a little used meaning in
trying to provide a concise definition.

Some of the wiki pages give an "OSM definition" that varies form the more
common and/or formal definition. In my view a weakness of OSM (but
sometimes necessary). The comment was made of the problem of definitions
that refer to outside sources that may change. I suspect when meaning
changes outside OSM, the new meaning is more likely to be used by new
taggers than the OSM definition. Users of OSM are unlikely to consult the
wiki, many will be unaware of the wiki, so will use current common meaning
for the tag. Language evolves, words come and go in popularity. Assuming
OSM should not also adapt will result in OSM maps that read for future
users like Chaucerian-English does now for current English readers (for
those unfamiliar with Chaucer'ian English, it can only be read currently by
experts or those with a dictionary). We do not want an OSM where as a
casual user you not only need a legend of tags, but a definition for each.

There is a bit of intransigence by some that limits changes that improve
OSM. I think management of change and consensus building will be important
to prevent (further) balkanization of OSM or it becoming irrelevant.

OSM also trains the repeat user, so the OSM conventions can not be ignored.
My expectation in searching OSM for a place that primarily sells
ready-to-eat food is it will be found under
amenity=restaurant,cafe,fast_food and I'll overlook businesses tagged
another way. If I'm visiting and you want me to patronize your ready-to-eat
seafood business, it should use one of the above amenity tags.

Murry
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tagging/attachments/20130918/fc517aa6/attachment-0001.html>


More information about the Tagging mailing list