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

[mail at marcos-martinez.net]

Hi all, 

one of the things that bothers me most of all in OSM is tagging
inconsistency, especially because we are talking about a database. This
thread shows a fundamental flaw in our processes: 

We have a relatively tedious proposal and voting process and after
finishing it turns out that the result... 

1. is ignored by those who don't like it 

2. is not implemented (as mentioned: it is a preset in JOSM) 

3. is unknown (I include myself; I wasn't aware that riverbank had been
downvoted in 2011 and have happily used riverbank, even "corrected"
water areas) 

I strongly believe that the correct way, once the voting has finalized a
standard process should be implemented: 

1. Data consumers/tools need to be informed about the change 

2. In case a mass edit is considered possible this should be performed
with due notification to data consumers/tools 

3. Mass edit should then be performed 

If the voting process is considered problematic then we need to adapt,
change and improve it but ignoring it only leads to further
fragmentation and ultimately anarchy. 

We really should make sure mass edits can be performed securely. I often
see debates that include the argument "yeah, this new tagging might be
better but it is a hell of a problem changing all the existing data".
This implies that we are NOT improving the data model and it will get
worse and worse over time. OSM will eventually become a huge
transatlantic oil  tanker incapable to change direction. This will be
particularly problematic when we REALLY have a new and much better
tagging scheme but are not capable of deploying it. 

Getting back to the concrete case here: I liked the wording "riverbank"
better because it was more concise, but this is just wording, so I will
use water=river as area from now on. For the same reason mass edit
should not be an issue I imagine but I don't have the background
knowledge to be sure. 

Cheers, 

Marcos Martinez 

Am 10.02.2021 15:43, schrieb Tomas Straupis:

> 2021-02-10, tr, 16:28 Marc_marc rašė: 
> 
>> this has already been voted (84% in favour) in 2011 [1]
>> the problem is therefore not the idea but the interminable transition
>> period and the inevitable "not of the same opinion as the result of the
>> vote)
> 
> The "voting" in wiki is totally non representative therefore it
> means absolutely nothing (or as we can see are damaging).
> Especially when we take into account that these proposals there
> created by people with 0 experience in fields that matter.
> 
> It is very sad and disturbing to see that OpenStreetMap fails to
> implement at least one of the safeguards against incompetent and
> destructive actions:
> a) require at least minimal experience
> b) provide hard limits on when some tagging can be changed or not
> 
> This leads to a classic situation of *** with a grenade decimating
> perfectly working parts of the system and there is no way to prevent
> that.
> Like "deprecate 100000's of usages of landuse=reservoir, deprecate
> 100000's of usage of waterway=riverbank, deprecate 100000's of usages
> of landuse=forest", what is next? highway=*?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tagging/attachments/20210210/22259cdf/attachment-0001.htm>