[Imports] Fwd: Re: Spanish cadastre

[Carlos Dávila]

El 28/03/12 15:40, Pieren escribió:
> I find a bit strange the way you plan the import. The source is
> providing the same dataset quality and completness for all
> municipalities but each contributor will be able to choose his own
> subset (e.g. with or without landuse or parcels) based on personnal
> assumptions. This will lead to a hugh inconsistency between
> neighborhoods.
>    
Should I stop mapping my city because another city a few kilometers away 
is much less mapped? If you are looking for consistency in a 
collaborative project where thousands of contributors add data I think 
you are in the wrong place. We've been asked to split the import by 
different type of data and now the problem seems to be right the 
opposite =-O
> You also have to know that once your tool is published, you will
> quickly find dumped osm extracts"ready to import" for the whole
> country even with parcels and landuse. And quickly some irresponsible
> contributors will blindly upload everything.
> > From our similar experience with the french cadastre, I would advice
> to publish a tool where the output is not configurable, data is
> already simplified and geometry issues fixed (like polygons, lines,
> points overlapping or not joining, etc). Do not expect that people
> will take care about quality before uploading. Even if most of the
> contributors will, some of them are just looking for quantity, not
> quality, as soon as it renders good enough on the mapnik slippy map.
> And they don't care if the map is not or hardly editable with current
> editors after their import.
I'm afraid idiots will ever exist, with or without cat2osm. Would you 
forbid cars all over the world because from time to time an stupid 
drives in a motorway the opposite direction just for jun? If you think 
this way, you should try to convince JOSM developers to disable WMS 
capability, because "some irresponsible contributors will blindly" use 
it to map over imagery not compatible with OSM license.
Carlos