[Tagging] Marking waterway=brook as deprecated and problematic

[Stefan Tauner]

On Wed, 30 Dec 2020 16:53:15 +0100
Martin Koppenhoefer <dieterdreist at gmail.com> wrote:

> Am Mi., 30. Dez. 2020 um 16:37 Uhr schrieb Stefan Tauner <
> stefan.tauner at gmx.at>:  
> 
> > On Wed, 30 Dec 2020 14:59:59 +0000
> > Philip Barnes <phil at trigpoint.me.uk> wrote:  
> >  > I have never used this tag however the wiki definition has widely
> > > missed the common usage of the word brook by a country mile.
> > >
> > > https://en.wikipedia.org/wiki/Rothley_Brook
> > >
> > > A brook as I understand it from growing up in areas where the term is
> > > commonly used, it is never intermittent.
> > >
> > > A brook is between a stream and a river. Usually too wide to jump but
> > > certainly too shallow or small for boats or to swim in. Deep enough to
> > > fall in and get soaked, something I used to do regularly as a child.  
> >
> > I think this is quite accurately captured by waterway=stream but we may
> > want to refine the definition regarding "jumpability" a bit.  
> 
>
> the given brook definition actually means that not a single waterway=stream
> would cover it, they would all be rivers. Only if the stream definition was
> "refined" aka completely changed they might cover it. I usually stop
> reading at the point where someone proposes to completely redefine an
> established and long standing definition for a major feature. It just isn't
> sufficiently realistic to merit a second thought, similar to your hiking
> club planning a hike to the moon ;-)
> Anyway, I invite you to have a second thought: how could we organize the
> review of 14 million waterway=stream objects plus 1.5 million rivers, that
> would have to be reviewed after we give up the main criterion for
> distinction and set a different one? What timeframe would be suitable? What
> are the benefits?

The example given (Rothley Brook) is tagged about half as stream and
half as river in OSM. Take a look at 52.6526663, -1.2048242 in Bing (and
probably other aerial services):
https://www.bing.com/maps?osid=b6417608-8f66-4c70-b87d-6186ade2b2c2&cp=52.652719~-1.205489&lvl=19&style=h&v=2&sV=2&form=S00027
That's about 1.5km away from where it "officially" changes to river in
OSM and there it's about 10m wide. Seems a bit too much to me for a
jump but maybe *my definitions are off* :P

Since sarcasm is often misunderstood I'd like to make more explicit
what I want to convey: The definitions in the wiki do not necessarily
reflect what people use if they "feel" they are wrong or inept - if they
are obeyed at all.
The "rule" we currently use is *very* subjective and definitely not
strictly obeyed. Thus describing my suggestion to refine (i.e., state
more precisely) what it means as "completely redefining" is a bit
ridiculous IMHO.

I would never tag a waterway that is only <20cm deep as river - no
matter how wide it is. This does not make any sense to me whatsoever
and I am convinced that the majority of mappers would not either.
The reason is rather simple: because such shallow but wide segments are
often short and are a matter of local geological or human influences.
It does not make sense to map a few dozen to maybe hundred meters as
river just because the ground is so sandy that the bank is very wide
and then change it back to stream downstream where it narrows again.

There is no perfect definition for the divide between big and small
rivers, not in OSM but also not outside of it AFAICT. Maybe academia has
some but I presume if they do at all then they are defined by flow
rate. Does anybody know?

-- 
Kind regards/Mit freundlichen Grüßen, Stefan Tauner