[Tagging] RFC proposed water property key 'ephemeral '

Warin 61sundowner at gmail.com
Fri May 25 22:03:28 UTC 2018

On 25/05/18 19:42, Christoph Hormann wrote:

> On Friday 25 May 2018, Warin wrote:
>> Intermittent does not equal ephemeral.
> So you have said repeatedly.
> I know this is hard for native English speakers but you need to accept
> that tags in OSM do not generally mean what the English language terms
> they use mean.

The English language has a lot of words, the choice should be made for matching the required meaning with a word.

> intermittent=yes in OSM currently means quite precisely what the OSM
> wiki says it means:
> "The intermittent=* key is used to indicate that a body of water does
> not permanently contain water."
> This includes at least a few 100k features you would define as
> ephemeral.  You can't change that with a proposal.
> There is a reason why such a broad and undifferentiated tag has become
> popular in OSM despite there being more differentiated concepts in
> existence:  Because more differentiation here is in many cases
> practically non-verifiable.  No matter if mappers observe on the ground
> or via images, the usual case is they see the waterbody in a dry state
> but see clear indications of recent water cover or water flow hence
> they can assume a non-permanent waterbody.  This is what you can
> currently indicate via intermittent=yes.
> But requiring mappers to guess what kind of time pattern the change from
> dry to water cover follows does not work.  Offering this as an option
> in case mappers have more in depth knowledge is a good idea, i said
> that in the past.  But making it mandatory is bound to fail.

I am trying to tightly define ephemeral so that it cannot be confused or used with with intermittent nor with seasonal.
If ephemeral is taken as being equal to intermittent than there is no point in having the key.
If a mapper cannot determine that something is ephemeral .... then don't use it! Simple.
Leave ephemeral for those that can determine it.
I do not require that all mappers use it.
That would be like having all mappers not use sport=multi but detail each sport using the ; as a separator.
There is no requirement to use or not use a tag. By providing the tag the data can have more detail when that detail is avalible.

More information about the Tagging mailing list