[OSM-talk] Licence of Facebook's derived road datasets? ODbL?

[Christoph Hormann]

On Friday 15 November 2019, Martin Koppenhoefer wrote:
> > > there isn't OSM data in their dataset.
> >
> > And neither is there is my ocean data set - the OSM data set used
> > only contains land masses, my resulting data set (D2 in Rory's
> > terms) only contains oceans.  So no OSM data in it.
>
> I doubt this cheap trick would pass when contested in a trial.

Well - it is not my cheap trick, it is facebook's cheap trick.  I am 
just following the lead here.  There is no principal difference between 
what facebook does and what my scenario describes.

> > If the question is not "addition or subtraction" consider the
> > following scenario.  You create a data set using some AI and big
> > data process of 'potential restaurants' world wide and create a set
> > intersection between those and the restuarants in OSM would the
> > results be a derivative of OSM data?
>
> yes, if you look at the intersection (data in both sets), it would
> be. If you took only what is not in OSM, I guess it wouldn't (no data
> from OSM contained).

So the set operation chosen (difference or intersection or any other) 
decides on the legal status of the resulting data set?

You are aware that a difference is the same as an intersection with the 
complement, i.e. A \setminus B = A \cap B^c - see:

https://en.wikipedia.org/wiki/Complement_(set_theory)

-- 
Christoph Hormann
http://www.imagico.de/