FYI: Plan9 open sourced

Gary W. Swearingen swear at attbi.com
Fri Jun 20 13:16:58 PDT 2003


Terry Lambert <tlambert2 at mindspring.com> writes:

> It is not clear to me that in (2)(a) "in source code and object
> code form" that the "and" could not be construed to require
> source distribution.

I'll admit that English has "and", "or", and "and/or" quite messed up
and often lacking clear meaning, strictly read, but I also have no doubt
that a court would think that giving someone license to do X and Y gives
them them license to do X, license to do Y, and license to do both, in
the absense of explicit language to the contrary.

In this case, (3)(A) seems to settle the matter: "Distributor may choose
to distribute the Program in any form...".

> It also seems that (3)(A)(b) might be construed as "viral", but
> it does not seem that it would be an offer of source.

That makes it no more "viral" than a BSD license or what copyright law
requires by default: that the licensor owns copyrights in the program
and its derivatives, even if the licensee also owns copyrights in a
derivative.  I don't see it as an attempt to assume ownership of
"contributions", at least based on (2)(a), though it comes close.

(The use of "tangible" in (3)(A)(b) seems wrong; I suspect that anything
distributable is considered "tangible" in legal jargon.  But IANAL.)

BUT other parts of the license appear (after a brief look) to make the
license worse than "viral": the broad and circular definitions seem to
make any derivative fall under the definition of "Program" and thus fall
under most terms of the license itself, removing any most choice from
the licensee about how he wishes to license his "contribution",
regardless of the other stuff about a "Distributor's own license
agreement", etc.  I find this overreaching to make the license highly
undesireable.

> The patent indeminification is something that most O.S. licenses
> forget to touch on.

I saw no specific patent indemnification.  (And this:

    The obligations in this section do not apply to any claims or Losses
    relating to any actual or alleged intellectual property
    infringement.

might even rule it out for the general indemnification mentioned.)

The patent-related terms seem over-reaching to me, along the lines of
saying "you loose your license if you use Windows".  It should be enough
for open source licenses to protect the licensed software from patent
claims, and should not try to effect "greater" goals, like effectively
preventing the enforcement of patents in derivatives, as this license
seem to want to do.  Software which restricts usage in such a manner,
even in a misguided effort to "keep it free", is no more open than the
design of a patented hardware widget, as it requires the "payment" of
something of value for it's use: a license fee for the widget and a
cross-license for the restricted-use software.

> The indemnification terms for "Commercial Distributors" vs. the
> controbutors will probably discourage commercial distribution,
> except by companies with deep pockets, or who are willing to
> accept potentially significant risk.

It kind of makes sense that a licensee's use (commercial distribution or
other!) of the "Program" shouldn't put the no-cost licensor at risk of
costs.  If the licensee doesn't accept the risks (to the licensor) of
his use, then the licensor must do so, and must do so before he even
knows who the licensees will be.  I agree that these clauses are onerous
to most open source licensees, but a whole lot of people are willing to
sign up for onerous licenses, as long as there's no immediate cost.

> All in all, this is in fact freer than the BSD license, from an
> other-than-commercial perspective, since they permit distribution
> of derivative works under the distributors license, so long as
> the conditions in (3)(A)(c)(i-iii) and (3)(B) involving additional
> warrants are complied with by the distributor.

See above for why that's way wrong, but in any case the BSDL also
permits distribution of derivative works under the distributor's
license, so long as the less restrictive (than (3)) conditions of the
BSDL are complied with.


More information about the freebsd-chat mailing list