[OSM-talk] Proposed mechanical edit - elimination of osmarender:nameDirection - blatant tagging for the renderer

Mateusz Konieczny matkoniecz at tutanota.com
Sat Mar 23 11:46:43 UTC 2019




Mar 23, 2019, 9:59 AM by simon at poole.ch:

>
> While there's a certain amount of unnecessary database bloat      caused by creating a new version, the real objection is that these      tags tend to survive in areas with low maintenance, aka low      amounts of gardening. Producing false updates (aka no real      content) just obscures that fact and makes it more difficult to      determine which areas need to be revisted.
>
>
It seems to me as not a real problem. There are many, many different indicators of such places
and automatic edits are suitable to remove only very small part of them.

I am pretty sure that after making all possible automatic edits it would be still easy to locate
such abandoned regions and human editing time could be used for things more useful than
manual removal of pointless tag polluting tag lists.

See for example http://osmose.openstreetmap.fr/en/map/ <http://osmose.openstreetmap.fr/en/map/>
(depending on location one may prefer different reports)

We have 800+ objects with source="Google Maps" (yes, it is on my todo  list, yes I notified 
DWG long time ago).
See http://overpass-turbo.eu/s/Hge <http://overpass-turbo.eu/s/Hge>

The bottleneck is not "we will run out of things to fix" or "we will run out of easy to find places
to repair", the bottleneck is time of mappers.

TIGER mess alone in USA is enough for decades or years of remapping, HOT low quality 
mapping in Africa or Haiti has massive amount of easy to find fixable data,
there are millions of instances of tags that has no reason to be present.

I see no point in wasting my time during normal editing on manual tag removal of keys
or tags that never, ever can be worth keeping and I would prefer to do it automatically.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk/attachments/20190323/57cc680c/attachment.html>


More information about the talk mailing list