[Tagging] Deprecation - waterway=riverbank vs water=river

Christoph Hormann osm at imagico.de
Sat Mar 6 12:35:50 UTC 2021

On Saturday 06 March 2021, Andy Townsend wrote:
> These last two are valid concerns, but aren't as "immediately
> requiring of DWG action" as if one of the items in the top list was a
> problem. The one that comes closest is (3) - someone (possibly Tomas)
> earlier said that they had or were aware of a data consumer that
> rendered "moving water" (rivers etc.) fundamentally differently from
> still water (e.g. lakes).  If that data consumer is "surprised" by
> "moving water" becoming "still water", then yes, they will have to do
> some work.

Please don't forget in your assessment the moving water features that 
are tagged natural=water without water=*.  My estimate is that there 
are probably more of those than what are tagged natural=water + 
water=river.  Those are the key to understanding the advantage of 
waterway=riverbank.  If you endorse use of natural=water + water=river 
for riverbank polygons you implicitly always state that the 
specification of water=river is optional and that tagging a riverbank 
polygon with natural=water only is valid tagging.

To be clear - in the past it has never been possible to distinguish 
between standing and moving water areas in OSM based on tagging with a 
really good level of consistency.  While before use of natural=water 
for riverbank polygons has started to be accepted this distinction has 
largely been followed in tagging it has never been done on a level of 
consistency that would have allowed high quality cartographic use.  
That is mostly the fault of map styles for a large part of OSMs history 
not providing any feedback on this distinction (and we still only have 
very few styles which actually do this).  However with the 
natural=water tagging for riverbanks widening in use the chances of 
this improving are decreasing fast.  The time window for OSM to retain 
the existing value in the data regarding the distinction between 
standing and flowing water and to maintain it on a level of quality 
that is useful for data users is closing fast.

If the trend from the past few years continues my estimate is that in 
2-3 years the quality of mapping in that regard has degraded so much 
that any data user in need of such information would come to the 
conclusion that it is economically advisable to generate this yourself 
from scatch instead of trying to work with the tagging in OSM.  For me 
personally this is a business opportunity - i can sell data users the 
service to supplement OSM waterbody data with information on flowing 
vs. standing water generated from remote sensing data and structural 
analysis.  For OSM however such a development would be a desaster - it 
would mean to simply give up on maintaining a distinction that is (a) 
trivial for mappers to record on the ground and (b) has historically 
already been recorded with a fairly decent level of consistency.

But i agree none of this is calling for DWG action.  Even if the DWG got 
active on this specific case it would not have a significant effect on 
the overall trends.  Where OSM develops w.r.t. waterbody mapping will 
ultimately be decided by how much ambition the OSM community 
collectively has in pursuing its goal to develop the best map of the 
world and how large on the other hand is the lure of retreating into 
the comfort zone of the mediocrity of doing good enough for the 
dominating use cases of today.

Christoph Hormann

More information about the Tagging mailing list