Re: Is this a free license?

[tb@becket.net (Thomas Bushnell, BSG)]

Russ Allbery <rra@stanford.edu> writes:

> So Dan certainly appears to believe that one can distribute the qmail
> source code along with a patch to that code and a script to compile it,
> based on his statements on his web pages.  That puts me back to trying to
> understand why I should believe one of you over the other.

I'm not particularly exercised about the case of qmail; I haven't
thought about it all that seriously.  It was brought up by someone
else who thought that if UnicodeData.txt was DFSG-free, then qmail
should be too.

The Dan-Bernstein-style reasons why qmail's license is unenforceable
(and thus qmail must be DFSG-free), if they are good reasons, are
entirely separate from that previous question.

I would note that patch files for software are generally thought to be
derivative works, and so I don't see any basis for his claim that they
are not.  (Note that the "patch files" for the Unicode data case are
rather different in the relevant respects from source code patches,
and so are not derivative works.)

I don't think Dan can rely on much precedent; the cases he's talking
about do *not* stand for the proposition that he goes on to talk about
later.

The point here is a very general one.  Yes, 17 USC 117 gives the owner
of a copy certain rights.

But that does not mean that you can therefore exercise those rights in
the free and clear.  Some exercises of those rights might run afoul of
the murder statutes.  Or the treason statutes.  (Does Dan think that
because you can distribute patch files, you must therefore be able to
distribute patch files that expose classified state secrets?)

So yes, you have the rights *granted by that section*, but only in
*general*, and not to the exclusion of *other* legal provisions.  

Your actions, in order to be legal, must *both* be legal in each step,
*and* be legal in the total effect of your actions.  And, dividing the
actions up between multiple cooperating parties doesn't change things
either.

So the cases that Dan is talking about apply *only* to the particular
steps, but whether the total effect is infringing is a *separate*
question.  In the Supreme Court cases he cites, there is no claim that
the total effect is infringing, but only the particular steps.  But in
the cases that apply for qmail (or KDE, or the GPL), it is the total
effect that matters.

Thomas