[linux-elitists] Re: [SPAM] Re: GPL Violations [was Re: Mobile Phone Choices]

Rick Moen rick@linuxmafia.com
Fri Jul 28 18:17:11 PDT 2006

[I didn't see Shlomi's post.  Maybe it had CCs?  If so, Mailman would
reject it as a linux-elitists post, as detailed on the listinfo page.]

Quoting Luis Rodrigo Gallardo Cruz (rodrigo@nul-unu.com):
> On Sat, Jul 29, 2006 at 12:38:22AM +0300, Shlomi Fish wrote:

> > Well, the reason for that is that MySQL wanted to sell a non-GPLed
> > commercial licence for its software which was generally GPLed. Now
> > since it used BDB, which was also pseduo-GPLed then MySQL had to pay
> > SleepyCat as part of the licence fees it recieved. (or Inno).
> > However, since Oracle has bought SleepyCat and Inno, it can now
> > refuse to exempt MySQL from the licensing of both the new and the
> > old versions of BDB or InnoBase.
> > 
> > Now, if BDB or InnoBase were BSD or LGPLed, then MySQL could have
> > continued to develop them on their own and use them. But since they
> > are GPLed - it's in a limbo.
> > 
> > Note that naturally, if you develop software under a GPL-compatible
> > license that links against GPLed code, then there isn't any problem.
> > However, that was not the case for MySQL.

There's nothing in copyright law[1] whatsoever that relies on "linking",
per se.  The determinative legal concept is, instead, that of
"derivative work".  (Be cautioned that that is a legal term of art,
i.e., subject to technical meaning[2] within the field of law, and if
you quote the definition in your dog-eared copy of Webster's Nth, His or
Her Honour will laugh at you.)

If you create a new work that is (in a copyright sense) derivative of
portions of Sleepycat's work that qualify for copyright protection, then
your work is encumbered by Sleepycat's copyright title, and you must be
careful to observe that firm's licence requirements concerning copying
of and creation of new works derivative of its -- and I use the term
advisedly -- property.

Concerning GPLed works, I keep hearing from computerists that the ruling
criterion is whether one incorporates literally copied code, or
alternatively "links to" it.  That is pure...  Comment dit-on en
anglais?  "Bubbe-myseh?"  ;->

[1] Relying on USA law primarily, but many other jurisdictions are

[2] There are standard conceptual tests applied by courts to determine
whether substantive literal or non-literal copying has occurred.  I 
have relevant links for relevant US Supreme Court rulings on 

More information about the linux-elitists mailing list