FC: John Palfrey's report from Boston in GPL MySQL license lawsuit

[Declan McCullagh <declan () well com>]

Previous Politech messages:

"Boston judge to hear first test of GNU license used in MySQL"
http://www.politechbot.com/p-03188.html

"NuSphere replies to Politech post about FSF and Boston lawsuit"
http://www.politechbot.com/p-03194.html

Background on the difference between the ordinary GPL and the Library GPL 
(aka Lesser GPL):
http://www.gnu.org/philosophy/why-not-lgpl.html

---

Date: Wed, 27 Feb 2002 13:49:43 -0800 (PST)
From: John Palfrey <jgpalfrey () yahoo com>
Subject: One perspective on Courtroom 13's Open Source story
To: declan () well com
Cc: ekelly () ropesgray com

Dear Declan:

I thought I would offer one person's take on what just transpired in 
Courtroom 13 at the US District Court in Boston, where Judge Patti Saris 
heard arguments in the complex dispute between MySQL and NuSphere/Progress, 
which centered in no small part on the use of open source software said to 
be governed by the GPL.

As previous posts to politechbot reported earlier, both sides had sued the 
other last summer.  At this stage, Judge Saris was considering whether to 
grant a preliminary injunction on two issues: 1) NuSphere's use of the 
MySQL-related trademarks and 2) NuSphere's use of certain code developed by 
MySQL and governed by the GPL as part of its Gemini product. Though she did 
not issue any orders on the spot, Judge Saris made it plain that she 
intended to issue a preliminary injunction against NuSphere's use of the 
marks in question (likely with 90 days to stop usage of the marks) and that 
she was not going to issue a preliminary injunction in the more complicated 
matter of the use of the MySQL code and terms the GPL.

1. The trademark issue was a fairly straight-forward contract construction 
discussion, with no cutting edge issues involved (from my perspective). She 
started with the presumption that irreparable harm existed, and never moved 
far off that point.

2. The harder issue, and the one on which she will almost certain not rule 
at this time, is the issue of whether NuSphere's use of the MySQL code 
under the GPL, and what NuSphere provides to end users pursuant to the GPL. 
One interesting point: it did not appear that anyone was arguing that the 
GPL did not apply or was not a valid license. It sounded as though the GPL 
was treated as any other license would be in a software context. Judge 
Saris seemed to focus on the question of whether Gemini constitutes an 
independent or a derivative work and whether the harm caused to MySQL met 
the irreparable standard. Experts -- none of whom were permitted to testify 
today, though Eben Moglen, perhaps among others, was in the room -- had 
filed what the Judge called "classic book-ends," or perfectly conflicting 
reports, on the question of the derivative work. Much of her questioning 
surrounded whether Gemini could operate without MySQL (as distributed, 
MySQL contends, it cannot) and whether or not the two products had been 
"integrated". She seemed to be moved by the NuSphere argument that there 
was no co-mingling of the source code and that "linking" to another program 
did not equate to creation of a derivative work. She also pushed hard on 
the questions of whether the distribution clause of the GPL was violated, 
though little progress was made on that point by either side. Ultimately, 
Judge Saris seemed unconvinced that MySQL could show a likelihood of 
success on the question of irreparable harm.

Judge Saris amused the crowd by referring to the open source movement as 
"like a religious movement." She also elicited snickers from certain 
members of the audience by referring to "Microsoft WordPerfect."

Before concluding the proceedings, she urged the parties to attempt 
settlement negotiations.

Best,

John P.

_______________________________

John G. Palfrey, Jr.
Ropes & Gray
One International Place
Boston, MA 02110
jpalfrey () ropesgray com
617.951.7037 direct dial
617.951.7050 facsimile

***********

Date: Wed, 27 Feb 2002 13:54:26 -0800
From: Craig Dickson <crdic () yahoo com>
To: Declan McCullagh <declan () well com>
Subject: Re: FC: NuSphere replies to Politech post about FSF and Boston lawsuit
In-Reply-To: <5.1.0.14.0.20020227125949.0247fa00 () mail well com>

Mark Lorion wrote:

> MySQL AB and the Free Software Foundation contends that
> NuSphere violated the GPL by simply linking proprietary software to the
> MySQL system using a public API. MySQL AB is interpreting the GPL so broadly
> that any commercial software that comes into contact with free software must
> also become free.  By that standard, a commercial email program would
> violate the GPL if it downloaded mail from a GPL-compliant mail server.

As someone who is neither associated with MySQL AB, NuSphere, or the
FSF, but who is familiar with both sides' statements on this issue and
the FSF's traditional interpretation of the GPL, I would like to comment
on this.

MySQL AB asserts that NuSphere's proprietary Gemini code is statically
linked to GPL'd MySQL code. As I understand the term "statically linked"
(I am a software engineer, so I expect I know what it means), this means
that Gemini and MySQL code are commingled in a single executable file,
and therefore together form a single program. This is clearly not the
same thing as an email client communicating with an email server; in
that case, two independent programs are involved. This is a distinction
that I have seen the FSF make many times, and it seems to me a very
sensible one, as it preserves the notion of a "software commons" without
making it impossible for proprietary software to interact with free
software. If it is true that Gemini code is statically linked to GPL'd
MySQL code, then Mr. Lorion's analogy is wildly inapt and displays a
lack of understanding of the GPL -- information about which is publicly
available on the FSF's web site.

Craig Dickson

***********

To: declan () well com
Subject: Re: FC: NuSphere replies to Politech post about FSF and Boston lawsuit
From: Jake Donham <jake () bitmechanic com>
Date: 27 Feb 2002 16:40:59 -0500

Hi Declan,

You'll probably get plenty of mail about this, but I wanted to give my
armchair lawyer's opinion that Mark Lorion seems to be misreading the
GPL when he says:

    Mark> MySQL AB and the Free Software Foundation contends that
    Mark> NuSphere violated the GPL by simply linking proprietary
    Mark> software to the MySQL system using a public API.

The applicable language in the GPL is

  These requirements [that derived works be licensed under the GPL]
  apply to the modified work as a whole. If identifiable sections of
  that work are not derived from the Program, and can be reasonably
  considered independent and separate works in themselves, then this
  License, and its terms, do not apply to those sections when you
  distribute them as separate works. But when you distribute the same
  sections as part of a whole which is a work based on the Program,
  the distribution of the whole must be on the terms of this License,
  whose permissions for other licensees extend to the entire whole,
  and thus to each and every part regardless of who wrote it.

The NuSphere code cannot be reasonably considered an independent and
separate work in itself in the absence of another implementation of
the MySQL APIs, because the code is useless without linking to
MySQL. If there were another implementation of the MySQL APIs,
NuSphere would still violate the GPL by distributing their code
together with MySQL.

The Lesser GPL exists for authors who want people to be able to link
non-GPL'd works to their code without infringement, but MySQL is not
distributed under the LGPL.

    Mark> MySQL AB is interpreting the GPL so broadly that any
    Mark> commercial software that comes into contact with free
    Mark> software must also become free.  By that standard, a
    Mark> commercial email program would violate the GPL if it
    Mark> downloaded mail from a GPL-compliant mail server.

A commercial email program can be reasonably considered an independent
and separate work because it also works with non-GPL'd mail
servers. Also, presumably the mail client and server are not
distributed together. "Coming into contact" is not a violation of the
GPL.

Jake

***********

Date: Wed, 27 Feb 2002 11:23:27 -0800 (PST)
From: Abraham Ingersoll <abe () dajoba com>
To: Declan McCullagh <declan () well com>
Subject: Re: FC: NuSphere replies to Politech post about FSF and Boston lawsuit

I'm a free-software publisher who has dealt with similar license issues,
even so far as bringing up concerns about the GPLv2 with Bradley
Kunz and in turn, Eben Moglen.

Just as Moglen's deposition concludes, I think it's pretty clear-cut
that NuSphere violated the GPL by not releasing the source when they
should. MysqlAB's not a happy camper from previous issues, and has
now decided to stick it to NuSphere with the GPL. Too bad for NuSphere.

As far as NuSphere crying about this setting a bad example for the free
software, and whining about it's implications in the business marketplace
-- SHUTTUP ALREADY!

NuSphere built something off of MysqlAB's back and then didn't play by
the rules (the GPL). Usually this wouldn't be too big of a problem, but
they had previously pissed off Monty (MysqlAB's lead) in many other
ways.

What's truly disgusting is how NuSphere's trying to muddy the waters with
claims about the GPL's supposed chilling-effect on intellectual property
rights.

That's just not the issue. The issue is whether they followed the GPL.
They did not. Now how exactly a judge is going to sort this out, and how
enforceable the GPL really is -- that's going to be fun to watch.
But NuSphere crying like a baby .. all that does is simply show what a
shallow corporation they really are.

NuSphere, bad. GPL, good. MysqlAB, dunno.

Abe

***********

From: Amos Satterlee <asatterlee () inta org>
To: "'declan () well com'" <declan () well com>
Subject: RE: NuSphere replies to Politech post about FSF and Boston lawsui
        t
Date: Wed, 27 Feb 2002 14:19:19 -0500

Declan:

I can't seem to find anything on NuSphere's original suit, just their
respones to the counter-suit.

The agreement that they (NuSphere) post on their site clearly states:

a. payment UP TO 2.5 million, based on coming to a second agreement. In
fact, they paid the money they said they would in the June 2000 agreement.

b. they were given right to register combination trademarks.

If their expert is correct and Gemini can act as a freestanding application
apart from MySQL, then I tend to agree that it's a stretch to GPL it -- or
are there other issues? It does seem to be a private spat wrapped up in the
flag of Free Software, but that's because I'm not clearly what NuSphere's
original beef was about.

Agreement: http://www.nusphere.com/misc_stuff/mysqlab_agreement.pdf
Expert: http://www.nusphere.com/misc_stuff/declarationofbrucefwebster.pdf

Amos

***********

From: "Vincent Penquerc'h" <Vincent.Penquerch () artworks co uk>
To: mlorion () nusphere com
Cc: "'declan () well com'" <declan () well com>
Subject: RE: NuSphere replies to Politech post about FSF and Boston lawsuit
Date: Wed, 27 Feb 2002 19:15:27 -0000

Hi,

> NuSphere violated the GPL by simply linking proprietary
> software to the
> MySQL system using a public API. MySQL AB is interpreting the
> GPL so broadly
> that any commercial software that comes into contact with
> free software must
> also become free.  By that standard, a commercial email program would
> violate the GPL if it downloaded mail from a GPL-compliant
> mail server.

The two cases you mention are wildly different in scope. In the
first case, the code linking against a GPL library leverages the
library in order to perform some work it could not do at all were
it not linked against it. In the second case, the email program
you mention merely *allows* one to use it alongside a GPLed mail
server. The email program would work without. Associating these
two wildly different cases shows either you don't know the obvious
difference between the two, or that you're trying to get people
who don't know the difference to believe that what you say is
true. I hope it's the former.
In the case of linking against a library, the case is very
specifically handled by the LGPL, which permits such linking.
The very existence of the LGPL (created, as the GPL, by the FSF)
shows that this distinction is not a fine one as you seem to
imply.

--
Vincent Penquerc'h

***********




-------------------------------------------------------------------------
POLITECH -- Declan McCullagh's politics and technology mailing list
You may redistribute this message freely if you include this notice.
Declan McCullagh's photographs are at http://www.mccullagh.org/
To subscribe to Politech: http://www.politechbot.com/info/subscribe.html
This message is archived at http://www.politechbot.com/
-------------------------------------------------------------------------