[Tagging] Definition of lake/pond as applied to stream/plunge pools

[Paul Allen]

On Wed, 23 Dec 2020 at 18:04, Brian M. Sperlongano <zelonewolf at gmail.com>
wrote:

>
> It seems like the convention for rivers is that the river's continuity
> (and name) are carried with the waterway=river ways and not the area
> polygons that cover the width of the river (regardless of whether you use
> the water=river or waterway=riverbank scheme).
>

So it appears to me.

  I also note that the distinction (effectively) between stream/river (the
> only variants that are in serious use) is that a stream is small enough
> that it's modeled by a way, while a river is large enough to require
> drawing a polygon.
>

Erm, nope.  The distinction is whether or not you can jump across it.
However,
wider rivers may benefit from a polygon.  But if you're in a hurry, or can't
be bothered, don't use a polygon.  Whenever I have masochistic urges
I extend the polygon on Afon Teifi further upstream.  This is as far as I've
gone from the estuary, and you can see that the rendering without the
polygon doesn't do it justice:
https://www.openstreetmap.org/?mlat=52.05829&mlon=-4.58272#map=17/52.05829/-4.58272
Follow the river a little way east from that point and you'll see a stream
join
the unpolygonned river.  Follow the river a little way north from that point
and you'll see the stream Nant Eifed join the polygonned river.


> I will make an assumption that there exists a class of these pools that
> are large enough that we would want to be able to map them as an area (and
> we wouldn't call them a pond),
>

Not so much wouldn't call them a pond as shouldn't call them a pond (unless
we declare pools to be honorary ponds for OSM purposes).

and there are also some that are small enough that mapping them as a node
> or linear way is fine also.
>

A node, maybe.  I'm not sure a linear way makes sense.


> 5. Be the same tagging for both rivers and streams
>

That could be hard.  It doesn't make sense to put a polygon on a
stream, they're not wide enough.

>
> I think it's fine if a river does not have continuity of water=river or
> waterway=riverbank polygons, as long as the waterway=river is properly
> contiguous.  I.e. I think it would be fine to have a sequence of river
> polygons with a stream pool polygon in between.
>

It would be rare (but not impossible) for the whole width of a river to be a
pool.  Mostly it's one side of the river where the flow rate slows.  But
since the
polygons overlay the water=river it should all work.  Maybe a carto problem
with name overlaps/priority, but that's probably soluble.

-- 
Paul
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tagging/attachments/20201223/ca3ab88c/attachment-0001.htm>