[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Linux mark extortion

On 6/17/05, Raul Miller <moth.debian@gmail.com> wrote:
> Linux already has already been established as meaning the linux kernel
> as well as distributions involving the linux kernel.  It's a generic
> term in that context.

That alone doesn't make it a generic term; it suggests that Linus
doesn't object to that degree of dilution through blurring, and
doesn't think those uses of his mark (including the broad outline of
distros' patch practices) cause confusion among customers for the
product he markets (the mainline kernel).

Have you, perhaps, case law to exhibit in support of your "generic
term" assertion?  It _might_ be a generic term by now, in some
jurisdictions, for operating systems built on minor variants of
Linus's kernel.  However, I certainly wouldn't want to be the first
person defending a trademark infringement or dilution case on that
basis in any given jurisdiction.

> That doesn't mean it's a generic term in all contexts -- there's no
> widely used Linux certification test suite, for example.

That's a case of "this product operates with the Linux program", which
is a usage frequently sanctioned by courts (as in the MySQL case) even
when the user is enjoined from otherwise using the mark in advertising
and marketing materials.  I think it quite improbable that the FooWare
Certification Test Suite (which Assists Developers in Verifying that
their Program Operates Properly on Linux) requires any trademark
license whatsoever.

Unless, of course, it results in dilution of a mark that is famous in
its niche; see Times Mirror Magazines v. Las Vegas Sports News at
http://laws.findlaw.com/3rd/991299.html .  Suppose, for instance,
Linus refused to accept a given multimedia-friendly preemption patch
into the mainline kernel because it makes one box in a thousand slag
itself, and someone started marketing a FooWare Multimedia Linux Test
Suite which implied that people should be using that patch on
multimedia-oriented systems.

Linus can probably say with a straight face that he "believe[s] that
there is a portion of the population that is adamantly opposed to
[machines slagging themselves] and that they would not look favorably
on any of [his] products if they thought [Linus was] promoting
[self-slagging machines] in any way."  (Quotation from Times Mirror,
mutatis mutandis.)  He can then point to the Times Mirror case in
support of a request for injunction against FooWare's otherwise
non-confusing use of the Linux trademark.

> Is there anything else that makes you think that this will be a
> concern for Debian?

That word "AUTHORIZED" is the big sticking point.  Is it correct to
say that Debian's use of the Linux trademark would make it appear to
an uninformed user that it has been authorized by the trademark

I think that, as far as a tarnishment theory of dilution is concerned,
all parts of the distro other than the kernel itself can probably
safely do anything with respect to the Linux mark that Connectix does
with the Playstation mark (see Sony v. Connectix).  But this doesn't
necessarily apply to the use of the Linux mark on a heavily patched
kernel, especially if the system doesn't really work right with a
mainline kernel.  As I understand it, Debian doesn't do so badly that
way at present (despite the use of cramfs initrds); but Montavista and
(if their current practice is anything like their behavior when last I
worked heavily with their releases) Red Hat might have cause for

On the other hand, Linux works so closely with Alan Cox that he
probably can't claim he hasn't more or less OK'ed Red Hat's
methodology all along.  Can Debian (or UserLinux or Montavista or
Ubuntu or whoever) argue similarly?  Debian may be more vulnerable
than some, since Linus almost certainly doesn't approve of a crusade
to expunge "non-free" firmware from Debian's kernels.  Time (and the
courts) will tell.

- Michael

Reply to: