> I have a problem that I would like y'all to consider.  A 
> company sells datasets.  They wish to cooperate with the 
> opensource folks to the extent that opensource folks can use 
> a free copy of their dataset. They would also like to be able 
> to sell that same dataset to people using proprietary 
> operating systems.  On the other hand, they would also like 
> to comply with the Open Source Definition.

A dataset is entitled to a copyright, albeit a thin one.  The individual
facts contained in it are possibly not copyrightable.  But the dataset
itself may be licensed on whatever terms the copyright owner wants, as a
copyrightable compilation, including under multiple licenses.  (You
omitted to tell us what the dataset contains.  The telephone book is
copyrightable, although the individual names and numbers aren't.)  

Suppose the copyright owner wants to let everyone use the dataset, but
any modifications that are made and distributed are to become available
under reciprocal terms (e.g., under the same license).  The OSL will
suffice.  I believe that a modified dataset is a derivative work of the
original dataset, but it might be best for your hypothetical company to
hire a lawyer to check out the cases on that point in his jurisdiction.
The issue is, if you merely add more non-copyrightable facts to a
dataset, is that considered a modification/derivative work?  By analogy,
if I were to take a published dictionary and add definitions and
republish it, am I creating a derivative work of the dictionary?  I
think so, but I'd want to verify this point.

The copyright owner can separately license the dataset to anyone he
pleases under any license he wants.  He can allow incorporation into
proprietary products without a reciprocity provision, for example, in
exchange for money.  He can promise follow-on versions (e.g.,
"maintenance" or "support") for a fee.  I've seen client companies make
good money selling compilations of facts gleaned from the public trade
press; their customers simply wanted to pay for the service of creating
and publishing the collection and keeping it current.  In another
example, companies can also make money selling collections of
freely-available clip art.  

To be sure that the dataset remains available under the open source
definition, your hypothetical company should license it under a license
that is OSI approved and that doesn't treat "software" as a narrow
category of intellectual property.  That's why I'd recommend the OSL or
the AFL.  The AFL, of course, doesn't include the reciprocity provision,
so it probably isn't what your hypothetical company would want.

> BTW, I just sent an announcement to 
> announcements@opensource.org.  If you're not subscribed, you 
> missed it.  :-)

I missed it.  :(  

/Larry

--
license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3