[Tagging] Separating usage docs from design docs (was: Increasing voting participation)

moltonel 3x Combo moltonel at gmail.com
Wed Mar 18 15:54:25 UTC 2015

On 18/03/2015, Martin Koppenhoefer <dieterdreist at gmail.com> wrote:
>> * key/tag pages would document the actual use (mainly observed via
>> taginfo)
> it is impossible to see  from taginfo what a tag is used for, and for what
> it can't be used. You only get statistics how much it is used
>> * key/tag pages could be kept up to date fairly objectively
> I find this difficult. If I start using a tag in the belief that it means a,
> and after two years people decide that this was a bad idea and now it should
> mean only a*, am I to review all my previous edits?

Yes, being objective and figuring out exactly what the current usage
is can be daunting, and taginfo is sometimes of little use
(landuse=forest vs natural=wood for example). But I think having a
stated goal of objectivity is still better than the current situation,
where some key pages document values that have never been used. Being
able to trust the content of a key/tag page without systematically
having to double-check taginfo and other sources would be a welcome

> Do we really need to change tag definitions, or would it be more sustainable
> to require new sub tags or alternative tags when the semantics should change
> or be amended?

We should certainly aim for backward compatibility when coming up with
new tags. It s not easy, we haven´t always succeeded. But that´s a
different topic.

>> * proposal voters should put the page on their watchlist, in case a
>> change in the proposal changes their opinion
> see previous comment

Yes, asking to watch pages is asking a lot. But I´d like to move away
from the formal drafted->proposed->accepted/rejected workflow, because
I think it just can´t work in OSM. That implies that proposals should
be able to evolve a bit over time. But if you make significant changes
after many people have voted, it´s probably better to create a new
proposal instead, to avoid backward-incompatibilities.

> also, I'd probably have to spend all day checking tag definition pages then

Not anymore than you watch actual OSM data, since tag definition pages
are supposed to reflect actual usage. So my suggestion should actually
reduce the need for page-watching compared to current workflow.

>> * proposals should only be "end-of-lifed" if there is near-unanimous
>> opposition and near-zero actual usage
> +1, if at all. Near zero usage should be <10

I don't like to give numerical thresholds, but yeah.

Another option for "end-of-lifeing" a proposal is if a newer proposal
replaces it in a backward-compatible maner.

More information about the Tagging mailing list