Re: GPL & Possible Derivative Work
On 6/17/05, Bruce Perens <email@example.com> wrote:
> Clean-room is an excellent strategy for avoiding copyright infringement.
> It enables someone to write a functionally compatible program without
> having seen the original. In an infringement case, it makes it entirely
> plausable to the court that the writer never saw the original
> implementation and was working only from the functional description.
> I'm not clear why you didn't consider that.
Plausible isn't much help here. Courts mostly only ask "is it
plausible that the resemblances are coincidental" when evaluating
claims of copying of non-literal elements from a rejected screenplay,
film scenario, creature sketch, etc. into a fictional work (such as a
movie) made later. That sort of "constructive availability" test can
weaken a claim of infringement at the factual copying stage, but
usually only if it's demonstrably improbable that the later author
even knew what was in the work alleged to have been copied.
When there's no question that the new work was written with knowledge
of the original's existence and general content, all that a
"clean-room" technique can do is help substantiate the defense that
certain non-literal similarities are a matter of engineering necessity
in the course of a Computer Associates v. Altai type
abstraction-filtration-comparison test. It is my impression that
(under US law) literal similarities are handled the same whether or
not they went through a "specification" stage, and are generally OK
only when they are demonstrably inseparable from ideas, methods of
operation, or compatibility requirements. I cite various examples
Note however that anti-reverse-engineering clauses are enforceable
(through a breach of contract claim), at least in negotiated
contracts, at least in some circuits; compare Bowers v. Baystate at
http://caselaw.lp.findlaw.com/data2/circs/Fed/011108v2.html . Note
that in that case the appeals court upheld the district court judge's
omission of damages for copyright infringement from the final
reckoning, on the grounds that awarding damages for both breach of the
anti-reverse-engineering clause and copyright infringement would
constitute double recovery.
As far as copyright is concerned, reverse engineering is reliably a
protected fair use in the US, except in the rare case (like Atari v.
Nintendo) in which an otherwise successful defense of fair use for
reverse engineering purposes is weakened by skulduggery such as fraud
on the copyright office. And even in that case, the court ruled that
the actual quantity of literal copying exceeded engineering necessity.
The criterion for copyright infringement in a final product is the
amount of copied material and the extent to which affirmative defenses
like copyright license and fair use apply to the product itself, not
to the process by which it was written.
Reverse engineering does frequently come up in cases where copyright
infringement is among the claims; see, for instance, Sony v. Connectix
(in which Connectix successfully justified its extensive use of the
Sony BIOS in the course of creating its PlayStation Emulator, partly
because Sony didn't claim to have found actual copying into the
emulator's final form) and Alcatel v. DGI (the one at
http://laws.findlaw.com/5th/9711339cv0v2.html ; there's a whole
DSC/Alcatel saga), in which the entire gamut of causes of action came
up except for claims of literal copying into DGI's product. But
"improper means" of reverse engineering are generally a trade secret
consideration, with little effect on the evidence for copyright
See also Bateman v. Mnemonics (at
http://www.law.emory.edu/11circuit/mar96/93-3234.ma2.html ; FindLaw's
copy appears to be mangled). The Bateman court vacated and remanded
the district court's judgment that copyright infringement had
occurred, partly because the jury had not been properly instructed
that compatibility requirements may justify literal copying. The
Eleventh Circuit said that it would be "an incorrect statement of the
law that interface specifications are not copyrightable as a matter of
law" but did "join [the 9th, 2nd, and 5th] circuits in finding that
external considerations such as compatibility may negate a finding of
The Bateman court's ruling was not long after Lotus v. Borland was
affirmed by an equally divided court, and other circuits weren't in a
big hurry to follow the First as far as "not copyrightable as a matter
of law". However, the Sixth Circuit went very nearly this far in the
recent Lexmark case, writing (in the course of reversing the district
court's ruling): "When a work itself constitutes merely an idea,
process or method of operation, or when any discernible expression is
inseparable from the idea itself, or when external factors dictate the
form of expression, copyright protection does not extend to the work."
It has been reported that the Supreme Court denied certioriari on
Lexmark, making it a pretty strong precedent.