[Tagging] Rivers classification

[Daniel Koć]

W dniu 06.08.2017 o 13:32, Richard pisze:
> as of rendering, respecting river width and doing something reasonable
> with intermittent flows would be a great progress.

What's the problem with intermittent flows? I'm not familiar with water 
tagging.

River width is a local property - it can vary lot in different places. I 
need simple, synthetic measure for lower zoom levels (like 
country/continent) without making any computations.

Stream order classification (or rather "waterway order" in our case) is 
very good for this, but it has few different versions, mainly (according 
to Wikipedia):

- classic stream order (also called Hack's stream order or Gravelius' 
stream order),
- Strahler stream order (also Strahler number or Horton–Strahler number),
- Shreve stream order (with Hodgkinson et al.)

but also topological stream order system and probably some other too.

We should allow any of them to be used, so a general namespace should be 
crafted. It could look like:

1. order:x
- order:classic (or order:hack or order:gravelius),
- order:strahler,
- order:shreve,
- order:topologic(al)

2. waterway:order:x

3. waterway_order:x (or stream_order:x or waterbody_order:x)

4. waterway:class:x (or waterway_class:x)

and possibly others. What do you think about it?

> Iirc the stream order issue has been brought up on some talk page
> previously.

Thanks, but it's just a brief mention:

https://wiki.openstreetmap.org/wiki/Talk:Key:waterway#Stream_Order

-- 
"Like a halo in reverse" [M. Gore]