New Open Source License: Single Supplier Open Source License

[Gary W. Swearingen]

Rahul Siddharthan <rsidd at online.fr> writes:
> Richard Schilling wrote:
>> The [L]GPL license makes this the submission of changes mandatory,
>
> No, it does not.  Please do your homework before posting such rubbish
> in public.  What is mandatory is distributing source code if you
> redistribute your binary.  Since you are not planning to allow
> redistribution, this is a non-issue.  The GPL and LGPL emphatically do
> not require submission of changes, private or public, back to the
> original author.  Private changes can remain in your hands, and you're
> required to give source code only to parties to whom you distribute a
> binary.

While it's probably low risk to assume that some private derivatives
are not subject to the "all third parties" clause, it's none-the-less
debatable, given the GPL's loose use of the word "distribute", which
doesn't have a well-known meaning, at least to typical licensors and
licensees who are the ones who are ostensibly using the language.  The
liberal interpretation of the GPL pretends that "distribute" and
"publish" are interchangeable and that the GPL's "distribute or
publish" is thus redundant.  Copyright law uses the phrase "distribute
to the public", which leads one to assume that there can be private
distribution too. E.g., distribution between two friends or two
companies or within a company or even between one person's two
computers.  It's not clear what the GPL means, even if the FSF has
made clear what they believe it to mean.  (The FSF isn't the only
user of the GPL, of course.)

Going further, I seriously think it's possible to read GPL clause 2
as an intertwined-rights license.  That is, it could be read as saying
that you may modify and distribute the derivative, but it doesn't
say you may modify and not distribute the derivative.  The subclauses
2a and 2c don't make much sense if you are not going to distribute
the derivative, and in clause 2c would actually be non-sensical on
two counts when it requires the startup message to say the the program
may be *re*distributed by users under the GPL, when the private
derivative has private non-GPL code and this is not under the GPL.

> GPL violation, there are legal recourses.  Sending the original
> author your private modification is not a requirement of any licence I
> know of (and is probably not enforceable even if you required it).

IIRC, it was on AT&T's Korn shell, and I'm fairly sure a couple of
others I've seen out of large companies.

> And your licence violates practically every other requirement of open
> source as commonly understood and as defined on www.opensource.org,
> starting with the very first (free redistribution).  So please don't
> call it open source.

Fooey.  Open source is software which you can read without cost.
Remember "The Open Software Foundation" and what licenses they had?
Don't let a Linux-heavy committee draw the line arbitrarily on the
wrong side of the GPL and redefine this term which has an obvious
natural English descriptive meaning.  (Thank the USPTO for disallowing
the trademark claim, for the same reason.)  Otherwise, I don't
consider GPL code open source, because it's not open to public
derivation without payment in the form of cross-licensing of the
deriver's work.  If you feel compelled to use the committee-approved
definition, please capitalize it.

> You're trying to hitch a free ride on the "open
> source" buzzword, without meeting its requirements.  The world does
> not owe you a living.

All true, but it's only a variant on what Perens and the
opensource.org crowd are doing with the "open" buzzword.  The world
does not owe opensource.org the world's copyright licenses, unless
they choose to accept their restrictive, proprietary licenses, with
the GPL being only one of the worst which meet opensource.org's
arbitrary definitions, created mainly so that "open" is more-or-less
equivalent to "GPL-compatible" so that the restrictive GPL can hitch a
free ride on the "open" buzzword.  At least Richard has the merit of
doing it for his living, and not for his pride, as is evidenced by the
GPL's prohibitions against putting a GPL'd "getline" call in an
otherwise BSDL'd program; that's about hubris, not freedom or
openness.

Have a nice day.  Way to go, JPL!