[OSM-talk] Import of Chicago Bicycle data.

[Jukka Rahkonen]

Russ Nelson <russ <at> cloudmade.com> writes:

> 
> As far as I know, we have no system in place for reporting diffs to a  
> data contributor.  We also have no system in place for merging in an  
> updated import.  If, on the other hand, he wanted to use OSM tools for  
> editing his data, we can output it as a shapefile back to him.  It's  
> really just a question of where he wants the canonical data to  
> reside.  If I had to guess, I'd say that he would be willing to give  
> us repeated data dumps, but what we do with them is our problem. Just  
> a guess.  I may be wrong.

In theory, all imported data should contain the import-ID and the unique ID that
is used by the original data provider.  That should be enough for identifying
the future updates.  Thus the login would be:
1) Update request is triggered by the data provider, and the message is that
Chicago Bicycle dataset feature with ID xxxyyyzzz should be updated
2)- If feature is unmodified (last edit made by the user who imported data, or
OSM changeset number is the original one) -> it can be updated automatically
  - If feature has been edited after import by OSM users -> mark it so that it
can be sent to manual check.

Unfortunately I really do not believe that automatic update can ever work with
road data because imported roads are topologically connected with other OSM
data.  Sending features to the update list might still work and be useful. 
Automatic updates should work with datasets which have self-contained features,
like building outlines or POIs.

-Jukka Rahkonen-