RE: [stellation-res] Stellation support for MySQL on Linux

["Mark C. Chu-Carroll" <mcc@xxxxxxxxxxxxxx>]

On Sat, 2002-11-02 at 20:08, Wynne Crisman wrote:
> I too am interested in using MySQL for a project I am preparing for and have
> noticed that the Driver is GPL.  If I were to write my own JDBC driver that
> communicated with the Database via IPC (sockets) would that require the
> MySQL folks to provide permission?  It would seem to me, based on my
> understanding of GPL, that communication with a separate process running GPL
> code does not in and of its self require that non-GPL process be released as
> GPL.  Please correct me if I am wrong so I don't go wasting my time.

One of the big problems with this kind of stuff is that
as non-lawyers, it's really hard for us to understand
this stuff in a definitive way. That's why I keep hedging
everything I say: I've spent a lot of time talking to lawyers, learning
what I'm allowed or not allowed to do;
and the biggest thing that I learned is that I really don't
have a clue when it comes to the legalities of this stuff.
The vocabulary is very specialized: words in legal documents don't
really mean what they mean in english. (Just like
we have our own vocabulary: talking about compiling
a program to a non-programmer, and they probably think
you mean something very different from what you actually mean.)

What you're essentially asking is: when does something
become a derivative work? If it's a derivative work, then
you're obligated to put it under the GPL. Exactly when
something becomes a derivative work is not precisely 
defined in the GPL, and the lawyers I spoke to told me
that the line where something becomes a derivative is
legally unclear.  So does using the wire protocol
make you a derived work? I don't know.

The folks behind MySQL have in the past allowed
people to develop non-GPL components that talk to MySQL
through the socket protocol. I don't know if that involved
negotiating special agreements to allow them to do that, or
if they consider the wire protocol open. 

> I also don't believe it is accurate to say that GPL is a PacMan like
> license.  Instead GPL is makes lots of business sense for any group or
> individual without the resources to protect their property because they
> don't have to protect their property under GPL.  GPL does allow for other
> licenses other than GPL, so an organization using any GPL library has three
> choices: release the application or library completely free under GPL, pay
> through the nose to obtain non-GPL licenses for the used GPL code, or remove
> all use of the GPL code (usually also expensive).  I would say that GPL is a
> business strategy, and so much a movement for freely available software or
> free expression of ideas.

I didn't intend to say that the GPL is a bad license - just
that it's a license that isn't compatible with our CPL
license. The GPL is a darned good license for many things. 
(There was a time when I nearly left IBM to go to
Ximian, and I would have been doing all of my work
under the GPL had that happened. And I think that Ximian's
reasons for using the GPL are sound.) 

As far as the "pac man" effect... As I said above, one
of the problems with it is that it isn't clear when something
becomes a derivative work in the eyes of the GPL. It's probably legal
paranoia, but without a clear definition,
the safest thing to do is to assume that *any* direct,
specific contact with GPL code is enough. And that's the
standard I was instructed to follow by the lawyers. And
like I said before - I'm the guy who'll get fired if I
screw this stuff up, so I'm justifiably paranoid. :-)

	-Mark



-- 
Mark Craig Chu-Carroll,  IBM T.J. Watson Research Center  
*** The Stellation project: Advanced SCM for Collaboration
***		http://www.eclipse.org/stellation
*** Work Email: mcc@xxxxxxxxxxxxxx  ------- Personal Email: markcc@xxxxxxxxxxx