[nycbug-talk] Sun News Roundup

[Miles Nordin]

>>>>> "ms" == Marc Spitzer <mspitzer at gmail.com> writes:

    ms> Was not most of the orignal GNU products rebranded BSD, and
    ms> BSD licenced, tools?

Maybe, but I wasn't aware of it.  I thought their project started by
working on some proprietary Unix platform, replacing the tools
one-by-one.  Do you have some citation?  Not that I necessarily agree
with you it's a problem or some kind of ``hypocricy'' if they did it,
though.

I just think you're slinging around a lot more than FUD to claim such
a thing if it's not true.  Trolling is one thing, but making up
garbage to get a rise out of someone is something else entirely, so I
hope you've some reason to believe it, other than you think it might
piss me off.  One can't prove a negative, so if this did happen then
the burden of proof is on you.  That said, here's the best I could do:

 http://git.savannah.gnu.org/cgit/bison.git/tree/REFERENCES

   describes differences between bison and yacc, but does not say one
   is based on code from the other.  It does say it uses a different
   algorithm suggesting it was at least rewritten.

 http://en.wikipedia.org/wiki/GNU_Compiler_Collection#History
 http://groups.google.com/group/comp.lang.misc/msg/32eda22392c20f98

   which does say they ``extended'' another compiler, but also says
   they rewrote it in C instead of some other Pascal-like language
   called Pastel.  Copyright protects the expression of ideas, not
   ideas themselves, so translating from one language into another
   should decisively remove any taint that may or may not have existed
   on the sources with which they started.

so, I'd say it looks pretty unlikely.

Another thing to keep in mind is that ``clean room'' techniques may
not have been well-known back then, because the problem of rewriting
source code that you already have, to change it from one license to
another, was probably not well-known---if you had the code, you kind
of just did what you liked with it because why-not-seems-reasonable,
which is how the AT&T lawsuit happened (CSRG and AT&T were both doing
huge amounts of it).

IIRC Stallman developed cleanroom techniques for reimplementing the
Lisp Machine work he was doing in crappy, free form (emacs).  so, if
anything he was kind of fucking up a LOT less than BSD people of the
time, giving a lot more respect to the specific copyright on the
source code he had in his hands.

    ms> There is no real way to argue that GPLing a BSD project is not
    ms> a violation of the authors intent.

so, first, for the _text_ of the license itself, obviously it's okay
with the BSD license to re-release as GPL provided you don't delete
the BSD restrictions.  BSD license does not have a problem with the
_additional restrictions_ imposed by the GPL, even though presumably
that's what you personally have a problem with.  While you have no
problem with releases where you get no source code at all---a view
which I still cannot understand.

The problem with living under both sets of restrictions concurrently,
is GPL will not allow you to redistribute a work subject to
4-clause-BSD + GPL at the same time because the advertising clause is
an ``additional restriction''---a work with both licenses stuck at the
top becomes as if the work had traditional
copyright-to-prevent-copying.  If the original holder of the copyright
offers it under dual-_either_, BSD or GPL, then a subsequent
redistributor could pick GPL, remove the ``additional restrictions''
along with the rest of the BSD license, and release as GPL-only---by
giving permission to subsequent redistributors to remove BSD whenever
they like iff they keep GPL, the GPL ``additional restrictions''
clause becomes happy.

But if the original holder released as 4-clause BSD only, as you
allege was the original source of most GNU tools including gcc, flex,
and bison, then of course BSD has a problem with deleting the BSD
license text from the redistributed copy altogether, because it says
in its text you can't.

You allege GNU simply removed the BSD copyright.  Why would they not
instead alter their license slightly to permit it to coexist with the
advertising clause?  Doing so wouldn't have compromised their goals
(in fact, they DID it with v3), and would have let them *legally* take
all the BSD work, slap GPL restrictions on it, and continue from
there.  They didn't.  I don't know why not.  But the fact that they
didn't suggests they didn't start with BSD source.

BUT you're asking me to defend from something slightly different, both
harder (because it's hazy) and easier (because one can't prove a
negative), than actual strict compliance with the text: you want the
_intent_ of the BSD developers.

I'm not sure I have to argue any of this as part of my beef with
Sun---there's no reason Sun and GNU can't both be wrong.  Sun doesn't
lose all their rights to complain about someone doing shifty things
with the CDDL like ``greenbytes'' might be, or Microsoft trying to
break Java in exactly the way Sun predicted they would and thus
violating the license, just because they're themselves being shifty
with the GPL.  irc arguing may work that way.  the Daily News may work
that way.  but the world doesn't.

Nevertheless I think it would be extremely easy to argue, starting
with the common intent-statement BSD-license authors make of, ``we
just want as many people to use our code as possible,'' and next, the
change to three-clause BSD license from the four-clause, the purpose
of which was to make it GPLv2 compatible by removing the ``additional
restriction'' the GPL couldn't accept---almost every BSD author
contactable agrees to 3-clause and thus implicitly to having their
work forked as GPL.  

I guess where your side goes from here, is to point at Theo and say
``choice of license doesn't indicate intent with a BSD developer,
because many of them (a) aren't capable of articulating their intent
clearly and consistently, and (b) tend to understand licenses rather
incompletely.  They say things like `I'm apolitical.  I just want to
get work done.'  ''  Well, obviously that situation is hard to even
comprehend, much less debate, but---if the author's intent is, ``I
want as many people as possible to use my work as long as they're not
Linux,'' AND if FSF really did start with BSD code which seems pretty
fucking unlikely, well then, I win again because Linux didn't exist
back then!

    ms> does not the FSF heavely and massivly misrepresent the word
    ms> free in their message?  GPL can not be free as the word is
    ms> defined in the dictionary,

their core message includes definition of exactly what they mean by
the word: the ``four essential freedoms.''  

 http://www.fsf.org/licensing/essays/free-sw.html

They even take pains to avoid the word free whenever practical because
of its multiple meanings: using it as ``software freedom'' or
replacing it with ``libre''.  If you go look at the actual message
coming out of FSF, I think it's virtually impossible to make the case
they're trying to profit from confusion.

If you're saying, ``I feel like I might have been confused and they
profited, for like two seconds, five years ago before I started hating
everything GPL, because here, look!, look at my dictionary,'' well in
that case I have to agree with you, that might have happened.

but IIRC my use of the word free was not the FSF's.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 304 bytes
Desc: not available
URL: <http://lists.nycbug.org/pipermail/talk/attachments/20090321/63af95b0/attachment.bin>