[Osmf-talk] Tag curation process (was: Toward resolution of controversies...)

Frederik Ramm frederik at remote.org
Tue Jun 9 07:47:33 UTC 2020


On 6/8/20 22:21, Craig Allan wrote:
> 2: A formal curation process for tags would be very welcome. I think the
> exploratory approach of 'any tags you like' has worked, but now has had
> its day. The issues with iD are largely an outcome of this problem. The
> next phase is to put in place a formal specification for tags serviced
> by a proper process to manage tag create-amend-delete events. I predict
> that editor and renderer developers will weep with joy if this is passed.

This is something people are often clamouring for but it comes with a
lot of problems below the water line.

One of these problems is that with the request for a definitive
catalogue of tags often comes the unspoken assumption that - because now
there is a canon of official tgs - everyone who processes OSM data will
evaluate these tags. For example, the official definition of
"access=private" will say "routing engines should not use this to
recommend a route" and everyone will assume that every routing engine
adheres to that. Which will never happen because we don't control the
myriad routing engines out there.

The longing for a simple world with clearly defined tags comes partly
from this assumption that the information transported in these tags will
then reliably be used (otherwise you could just make something up, or
use the "description" tag to record the information you want to record).
But a clearly defined tag catalogue will never deliver that.

I believe that a formal curation process for tags will only
superficially make things easier, and simply move problems elsewhere. At
the very least, it needs to be extremely well considered.


Frederik Ramm  ##  eMail frederik at remote.org  ##  N49°00'09" E008°23'33"

More information about the osmf-talk mailing list