Re: Responding to your CDDB Inquiries (fwd)

Greg Stein (gstein nospam at lyra.org)
Tue, 09 Mar 1999 12:46:10 -0800

[copying the cdin nospam at cdin.org mailing list to get this into a web archive
for future readers]

Kyle R. Rose wrote:
>
> > Taking the hint from something I saw in /. today, we should really
> > discuss a license for the data we are about to gather, in order to
> > prevent more harming to the community.
>
> I have serious doubts about the legality of GPLing a protocol; nor do
> I think we should. I'll get to the actual data in a moment.

I don't think you can. You can patent techniques used to drive a
protocol, but you can't patent the protocol itself. Copyrights don't
cover protocols, so that means that protocols are effectively public
domain.
(they can be regarded as trade secrets, but once the cat is out of the
bag...)

> As I said yesterday, the library code should be distributed under a
> non-viral license. That means LGPL at the very least, up to BSD minus
> the advertisement clause at the most ideal. Remember that a lot of
> players are not open source, but we still want them to adopt this
> protocol. It's the lesser of two evils, right?

Absolutely.

People have mentioned the LGPL. That isn't good enough. It means that
you still must distribute your code in a linkable form such that a
person can relink with a modified library. That won't work for most
people.

It must be BSD-like.

> As for the data... well, arguably, the CDDB entries aren't Escient's
> property anyway, as they don't own the copyrights to the album or song
> titles. What they can copyright is the _anthology_, i.e., the
> formatted collection that makes up the CDDB.

Whatever they want to claim for new entries, the old set is certainly
available. I've loaded about 50k of them into my server at
http://www.cdin.org/

> So, ideally, what we need is a copyleft license which mandates that
> any changes to the database be made publicly available.

This would be nice, but I'm not sure that you can do this for data.

As I've mentioned before, the data is public information (subject to
"fair use" copyrights). A compilation can be copyrighted in a specific
compilation; once you extract and reformat, though, it is yours to use.
So... even if we attempted to protect the data, it is still usable my
anybody in any form they please.

Luckily, I believe this gives us our edge against the CDDB database.

Our best protection is to just keep copies distributed as much as
possible.

Cheers,
-g

--
Greg Stein, http://www.lyra.org/