[OSM-talk] How to correct logical errors in some effective way?

[Ulf Lamping]

Rahkonen Jukka schrieb:
> Hi,
>
> When I am looking at the Scandinavian data loaded into PostGIS database with the fine osm2pgsql utility (thanks, Artem) I can simply find hundreds or thousands of logical errors in data.  For example this is an excerpt from the list of highway tags used in the data:
> ...
> residetial
> roundabout
> secondary
> Secondary
> secondary link
> secondary_link
> service
> Service
> services
> servie
> servoce
> sevice
> ski_jump
> Skolestræde
> snowmobileway
> sservice
> stairs
> step
> steps
> Steps
> tertiart
> ...
> There are lots of pure typos which could be most effectively corrected directly in the database without using JOSM or something in between.  Is there somebody whom I could send a list of the items that need update?  I can get osm_id of those features, I suppose it is unique and usable for batch updates.
>   
I've done some manual typo cleanup based on the german tagwatch data 
some weeks ago, but that's far too annoying to do it on a regular basis. 
BTW: it's funny to see how many typos you can make in amenity=restaurant ;-)

While we should be careful about converting stuff automatically (mappers 
may just disagree in the changes - german street name Straße vs. Strasse 
is such a case), I perfectly agree it's a good idea to correct obvious 
typos in the tagging.

A simple approach like having a list of rules like "highway=sevice must 
be corrected to highway=service" would be good enough here IMHO. If 
anyone comes up with a solution for this typo correction thing (I'm not 
a database guy), I would love to join forces by looking at the tagwatch 
lists and would help to maintain such an automated "typo correction list".

Regards, ULFL