[Tagging] motorcycle:scale

[Kevin Kenny]

On Tue, Feb 5, 2019 at 2:20 PM Martin Koppenhoefer
<dieterdreist at gmail.com> wrote:
> Indeed I believe we shouldn’t create feature pages for “in use” features until they become de-facto. In use means at least one occurrence.
>
> There is also “rejected” among the possible status values, I don’t think you suggest to create feature pages for rejected proposals?


As long as we indicate the status on the page, I have no trouble with
having feature pages for all of the above.

Consider the following scenario.

Mapper Mary sees a set of features in the field: call them 'betas'.
She creates a small set of OSM objects that are tagged 'alpha=beta',
because her Google-fu has failed her when trying to find any existing
tagging for betas on the map.  Following good practice, she
immediately creates a Wiki page indicating what the tag means, and
indicates that the tag is in use but not approved by the community.

Once she's got some experience with the tag (and perhaps Mapper Mike
joins her in mapping betas), she decides that betas may be of more
general interest, and floats a proposal.

The tagging mailing list discusses six alternative ways to represent
the proposed feature. Someone argues that betas don't actually exist.
Someone else insists that betas are really just another kind of bees.
A third person insists that there's no way to be certain in the field
that what you're looking at is a beta, so the proposal should fail on
verifiability. Another insists that before betas can be mapped, we
need a whole taxonomy of Greek letters, and the discussion rapidly
devolves into a long digression about whether θ and ϑ are really the
same letter because one is printed and the other is cursive.

The proposal is rejected because all the proponents of the
alternatives vote "no". Note that none of the alternatives would be
accepted as a proposal either, since, as is usual, no consensus is
achieved as to what alternative tagging should be used for the betas
that Mary and Mike have mapped. The result is that now there's a
rejected proposal without a good alternative.

What happens next? The most likely outcomes:

1. Mary and Mike remove the objects that they created, since there's
no appropriate way to tag them. They decide to do their mapping atop
some commercial platform. OSM loses them as mappers, and loses the
objects that they created.

2. Mary and Mike leave the objects in place with their invented
tagging, but remove the Wiki page that describes the tag because the
proposal was rejected. The rejected proposal page is eventually
proposed for cleanup, and winds up being removed. The objects are
'obviously' incorrectly tagged, and some other mapper, not knowing
what the situation is, either mistags or removes them.

3. Mary and Mike leave the objects in place with their invented
tagging, adding a note to the Wiki page describing the tag that a
proposal for the tag was rejected without a consensus as to
alternative tagging. Other mappers on encountering the object can look
up the 'alpha=beta' tag and see what Mary intended when she added the
objects. Perhaps someday the tag will ascend to _de facto_ status, or
someone else will manage to be persuasive enough that the proposal
will succeed.

This is a scenario in which, as far as I can tell, Mary and Mike have
done everything right, but the community has failed them. Which of
these likely outcomes is the 'least worst'?

We have a formal rule of "any tags you like", but apparently we're
converging on a _de facto_ rule of "ask permission before applying any
tag unless you see that it was the subject of a previously approved
proposal." Is that the rule we want?