> From: [EMAIL PROTECTED]
> Subject: Re: Yikes - it's opensauce
>
>
> Bernd Kreimeier wrote:
> >
> > John Keiser writes:
> > > What we're left with is this: the original *program* is freely
> > > redistributable, but a *derivative work* is freely redistributable
> > > under two conditions: it must contain 60% or more of the original
> > > code, and it must be a JDK 1.1 compiler.
> >
> > There are days when I think that, despite me freezing
> > and starving, RMS is just a 100% right.
>
> [disclaimer: IANAL, and I haven't scrutinized the license in detail]
>
> A consensus seems to have been reached on slashdot (at least, as much of
> a consensus as is *ever* acheived on slashdot) that the 60% clause only
> applies to patents. The definition of "Program" includes modifications
> made by yourself, so you *ARE* free to distribute modified versions
> without restriction.
The only place it grants the right to distribute without restriction is in
paragraph 1, section 2, but it qualifies "Program" with the phrase "as
distributed by IBM." Additionally, it gives you the right to prepare
modified versions of the code in that paragraph (to tinker), but not to
distribute those modified versions. So the first paragraph hasn't
guaranteed anything except that the source code is freely available to all
and that you can legally modify it.
> IN ADDITION, IF your modified code (1) is a Java
> compiler according to the JLS published by Sun, AND (2) includes 60% of
> the original code (whatever that means), you have rights to any IBM
> patents inherent in the Jikes code.
>
That portion of the code, listing those qualifications, is the only place
that also grants the right to distribute modified code. Thus the
distribution of modified versions is inexorably tied to 60% and JDK1.1
compiler clauses.
> Even the GPL doesn't grant patent rights!
>
I doubt GPL'd code has ever had patents in it. This is a move of necessity
if one is to open up patented source code.
> This is a good license in almost all respects, AFAICT. The only
> "non-free" thing about it is the termination clause... which I hope
> enough people will point out to IBM to make them change it. Free
> software should not be able to be unfreed, ever, no matter what.
>
These are the facts: there is language in there that *does* grant patent
rights, in limited scope, but it also grants distribution rights in that
same limited scope. No clause except the patent one grants distribution
rights.
Bruce Perens of Open Source appears to agree, and he told me in email that
Open Source will announce that the license is not open source, but that they
are working with IBM on it. Further, he said the problems are in #2 (the
grant of rights, the place I have been talking about) and in the termination
clause.
This is an important issue. The restriction of distribution to the source
of the program means that they *retain legal control of it*. Open source
projects cannot be truly open if the company that opens them retains legal
control.
--John Keiser
