[OSM-legal-talk] Database and its contents

Rob Myers rob at robmyers.org
Mon Nov 22 22:55:44 GMT 2010


On 11/22/2010 10:25 PM, Ed Avis wrote:
> 80n<80n80n at ...>  writes:
>
>> The relationship between ODbL and DbCL is not very clear and I'm not convinced
>> that lawyers really understand the distinction between a database and it's
>> content.  I'm certain that it isn't understood by most ordinary people.
>
> I work with databases every day and I don't understand how the 'database' versus
> 'contents' distinction is meant to apply to maps and to OSM in particular.

Imagine a database of names, song titles, photographs, recipes, poems or 
credit card numbers.

The individual names, song titles, photographs, recipes, poems or credit 
card numbers are the content.

The database of names, song titles, photographs, recipes, poems or 
credit card numbers is the database.

Some of those kinds of contents can be individually copyrighted (notably 
photographs and poems). Some cannot (notably recipes and credit card 
numbers).

You can claim a copyright on a collection of works that isn't a 
database. It's how public domain clip art collections, public domain 
music CDs, and public domain text reprints end up with copyright notices 
on them.

The copyright on the collection means that people can use the individual 
contents of the collection, but if they reproduce the collection as a 
whole they run up against the collective copyright. At which point 
between the entire collection and a single item do you stop infringing 
the copyright on the collection if you copy it? Ask a lawyer...

Copyright (and DB right) on databases and their contents is, to my mind, 
similar to this. It's copyright on the collection, not the collected.

What seems to throw people when we are talking about geodata in a 
database rather than a collection of poems/photos/songs is the 
granularity of the contents. But it doesn't really matter whether we 
regard points, ways, uploads or any other unit as the content of the 
database. The content of the database is any pieces of data smaller than 
the entire database.

The fact that the contents is DbCL is what allows you to make BY-SA maps 
from the database contents at all (thank you Anthony for helping me to 
understand the details of this). The fact that the database is ODbL is 
what ensures that substantial extracts of that database (substantial 
collections of that contents) cannot be used without passing on the 
freedoms that the ODbL gives.

If that's too much, think of the database as the totality of data plus 
the DB schema, and think of the contents as anything less, and don't 
worry about the precise point at which the land meets the sea. ;-)

Any complexity in this is a product of the law not the licence...

(I am not a lawyer, this is not legal advice. If I'm wrong about any  of 
this *please* correct me. :-) )

- Rob.



More information about the legal-talk mailing list