[Tagging] Rivers classification

Wolfgang Zenker wolfgang at lyxys.ka.sub.org
Mon Aug 7 19:59:26 UTC 2017

* Mark Wagner <mark+osm at carnildo.com> [170807 20:45]:
> On Mon, 7 Aug 2017 16:37:52 +0200
> Martin Koppenhoefer <dieterdreist at gmail.com> wrote:
>>> On 7. Aug 2017, at 10:51, Frederik Ramm <frederik at remote.org> wrote:
>>> I agree that some sort of river classification might be helpful but
>>> you cannot expect a mapper standing before a river to first analyse
>>> a large dataset before they can find the right classification tag -
>>> that would totally run counter of "on the ground verification".  

>> I don't buy this argument because the situation for roads is the same
>> and we do expect from mappers to analyze the network.

> At least in developed countries, you can get an idea of a road's
> classification in the network just by looking at how it's constructed.
> [..]

> You can't do the same with rivers.  The Clark Fork River and
> the Colorado River have similar average flows, but the Colorado would
> have a higher classification by any of the proposed measures.

The Colorado River might have a higher classification by any of
the proposed measures, but the question is if the classification
would be useful to us. Regarding rendering I think that any river
with some degree of significance will be mapped with riverbank polygons
eventually. The renderer can use those to see if rendering at a given
zoom level makes sense. For rivers and streams of lesser significance,
counting of tributaries and looking at width tags could be used, all
without having to add some sort of classification to the data manually.


More information about the Tagging mailing list