Where do I begin?

Tony Kimball alk at pobox.com
Mon Mar 9 12:25:38 PST 1998

Quoth John D. Gwinner on Mon, 9 March:
: However, by maintaining clean room, then it's easy for Kaffe to say "Not an
: intentional violation, we haven't seen the source; it's a parallel
: development"
: Do you see now what I'm saying?

Certainly I understand the point that it is easier to justify a
positive claim that no violation has occured when not only the
evidence but also an a prioristic argument on the basis of a
presumption that none of the contributors have falsely claimed to 
be virgin.  

I do not regard this as a serious issue for kaffe qua public project, but
I opine that anyone who wishes to compete against JavaSoft using kaffe
has it in their interest to preclude effective suit to a much greater
degree than does the project per se:  Were the project open to
non-virgin contributors, a risk would arise that some contributor
would violate the copyright, thus exposing the commercializing agency.

: In the courtroom, it's not the absolute of *IF* something has happened, it's
: what you can prove.  

And indeed the burden of proof lies upon the plantiff.  I understand
that the plantiff must first prove that the opportunity of plagiarism 
existed before being able to prove that plagiarism occured.  However,
this necessary condition is by no means sufficient.

: With a clean room, Kaffe can *prove* it hasn't seen the
: source, therefore can't have infringed.  

Only to the degree that the contributor claims of virginity can be
proven.  If the pool of actual contributors is large enough,
it is very likely that some one of those contributors is not in
fact virgin.  For kaffe, I suspect the pool has been small enough
to prevent such an occurence.  But it is precisely that small 
contributor base which I bemoan, and would like to see addressed.

: Without a clean room, Kaffe can't
: prove anything, and it's easy for Sun to prove that the source code is the
: same (at least in portions).

I disagree, in as much as the clean-room claim is difficult to prove.
I would agree with a more moderate claim that Sun (really, JavaSoft)
would find it infeasible to prove plagiarism if it could not prove
that the clean-room claim was erroneous.

My position is that the benefit to kaffe per se would be greater if
the clean-room condition was abandoned.  The point is arguable, in as
much as the commercializers (Cygnus, in particular) who are the
current leaders in contribution to the project would be faced with a
marginally greater risk of suit and hence might concievably elect to
fork.  However, even if a fork occurred, were an open repository
maintained by an independent agency, it could freely incorporate any
and all Cygnus code, just as Cygnus could incorporate whatever code
from that repository they could satisfy themselves as being untainted.
Moreover, the pool of potential contributors would be increased, I
estimate, by at least an order of magnitude!  (My blind guess is that
the overwhelming majority of parties sufficiently interested in Java
to contribute have also been sufficiently interested to download a

The point may well, however, be moot: For one thing, my own resources
are insufficient to maintain a fork on my own initiative.  For
another, it remains to be seen that the Transvirtual licensing terms
will prove unacceptable to open source advocates.  Finally, the
incentive to violate the clean-room condition and maintain a fork
under those terms does not exist until and unless the currently
inactive pool of potential tainted contributors were to become active.
It is conceivable, however, that were such a pool to manifest itself,
some agency with greater resources than my own might volunteer to
maintain an open repository.  W3C comes to mind.

If nothing else this thread has hopefully served to publicize the
fact that the intellectual property issues involved are restricted to
those of copyright, and do not extend, e.g. to trade secrets.

More information about the kaffe mailing list