[OSM-talk] Overlapping ways

SomeoneElse lists at mail.atownsend.org.uk
Wed Jan 9 21:31:57 GMT 2013


Rob Nickerson wrote:
>
> Am I doing something wrong? Is the fixer tool flagging something up 
> incorrectly?

In cases such as this I normally say to the other mapper that I was last 
there on so-and-so date, and when I was last there it looked like X; and 
ask whether perhaps he's been there more recently and the geometry's 
changed?

Personally I believe that the various QA tools do an excellent job, but 
it'd be simply impossible to miss all the false negatives and not catch 
some false positives - it has to be up to the mapper to decide what's a 
real issue and what's not.

(begin rant)

I wish more of the people doing these remote corrections would actually 
talk to the person who did the original mapping in the first place.  
Perhaps the way was drawn by a new mapper who actually has lots of 
questions about how to do things, but doesn't know who or where to ask.  
Maybe it's a mistake by someone who's been mapping for a while (we all 
still make them!), in which the best person to correct the error is 
surely a person who's been there rather than a person who hasn't.

In some cases it does make sense to correct remotely (perhaps 
non-connecting footpaths that match GPS traces that were drawn by a 
mapper who hasn't been seen since 2009 would be an example), but in many 
cases I would argue that it doesn't.

(end rant)

As an interesting aside, what I've found myself doing more frequently 
recently is revisiting places that I'd mapped previously that had been 
subsequently "armchaired".  In almost all cases what resulted from a 
resurvey wasn't exactly the same as from the original, but slightly more 
nuanced and with a lot more detail - revisiting isn't necessarily a bad 
thing.  Still, it can be annoying to have to go back and resurvey an 
area because someone has "corrected" it to look like an old Bing photo, 
prompted by a false positive on a QA site.

Cheers,
Andy




More information about the talk mailing list