Re: [Gnu-arch-users] [OT] Debian should move all GPL sw to non-free

[James Blackwell]

On Sun, Mar 28, 2004 at 04:58:06PM +0800, Cameron Patrick wrote:
> Yikes, I seem to have touched a raw nerve there... Sorry.  I actually
> /agree/ with your conclusion that Tom's argument is invalid; it's just
> your argument for why he's wrong that I'm not so sure about.  I certainly
> wasn't intending to flame you or attack you personally.

No problem. Appologies from me to you as well. 

I'm working from a difficult position. Toms argument is pretty weak to
treat properly. Though he makes some pretty clear statements throughout,
he doesn't tie them together as well as he should have, which makes it
difficult to treat in a philosophical way.

If you agree that I've knocked out his supports (namely that the GPL must
be treated the same as the GFDL), does his argument still stand? 

If not, then I've done my job. Remember: my job isn't to prove
him wrong; rather its to prove that he didn't prove himself right.

Can you provably state that his article still proves itself?


> On Sun, Mar 28, 2004 at 02:53:35AM -0500, James Blackwell wrote:
> 
> | > Hate to say it, but Tom's rant was actually a good deal more coherent
> | > and better written than yours.
> | 
> | Beg your pardon? I think I did a pretty good job. Did you read my
> | essay?
> 
> Yes, of course I did.  In retrospect I might have been a bit negative
> towards it, but I stand by the gist of what I said.  As I mentioned
> above, I agree with your conclusion about the GPL vs GFDL, it's just
> that I didn't find your essay particularly convincing.

Well, I think that what Tom is trying to do is tie the GPL and GFDL
together. In regards to the DFSG, either they're both free or they're not.

> | > Tom's argument can, I think be expressed more succinctly as:
> | >
> | > 1) To be DFSG-free, a licence must allow binaries built from arbitrarily
> | >    modified copies of the source code
> | 
> | Yes. I'm not aware of anybody arguing that point.
> 
> Okay.  I think that rejecting this point is the only sane way out
> here.  The DFSG, at least as I interpret it, is actually ambiguous
> about whether anything which allows some forms of modification is
> sufficiently free, or whether something must allow arbitrary
> modification to be free.  In practice, it's somewhere in between.
> There is no precisely defined point after which restrictions on
> modification are considered to be non-free; it depends on the person
> interpreting the guidelines, and how it fits into the licence as a
> whole.  It's a difference in degree, not a difference in kind, though.
> 
> As you yourself point out, the important difference between the GFDL's
> invariant sections and the GPL's 2c is that invariant sections can be
> of any length, on almost any subject, and can't under any
> circumstances be removed or altered.  The GPL's restriction only
> requires a short message (and even then only printed under some, very
> specific, circumstances), with clearly defined content.  The message
> itself can even be modified so long as it fulfills some criteria.
> /That/ is why GPL 2c is (grudgingly) considered to be free, whereas
> the GFDL is not.

You're being pretty fair here. I suppose that I could have attacked point
1 as well. 

The way that I chose to refute Tom's point was along these two lines: 

1. The GPL and the GFDL are not identical
2. We are thus allowed to keep one and throw the other out

I did this by first showing that they aren't the same (though I do leave
room for him to argue that they are graduations of the same thing).
Because I show that the GPL is not the same as the GFDL from a liberty
stanpoint, this leaves him with three possible positions:

1. If Tom sticks to black and white, the GPL must be free because the 
GFDL takes the position of nonfree. Here, his argument is lost because
he is now reliant upon a logical fallacy.

2. He can acknowledge the GPL is different from the GFDL, and his argument
looses on its merits

3. He can start over with a different argument, claiming that while the
GPL is not the same as the GFDL, it is similiar enough to qualify. 

#3 is a possible avenue for him, but not an easy road to take. In order to
go this route, he has to account for the later writings from the Debian
team which clarify their position. 


> So while neither pass a strict "must allow arbitrary modifications"
> check, in practice this doesn't matter, because Debian /doesn't/
> require this.  What I think Tom was trying to argue in a different
> message is that if Debian doesn't require /all/ kinds of modifications
> to be allowed, then restricting any subset of modifications must be
> okay, and therefore if the GPL is free then the GFDL must be free too.
> This, to me, seems like a false dichotomy.
 
I have no doubt that he'll probably use that argument next, but its not
the argument he's making here. When he makes that argument, I'll probably
take that one one too

But I have to treat his ideas here, in this essay, on their own merits. I
have to provide a resonable defense on his behalf. But I can only go so
far without risking a straw man attack.


> | > 2) GPL 2(c) disallows a particular type of modification to the source code
> | 
> | Yes. However, Tom himself argues that those are not source code, but
> | political statements.
> 
> I saw that Tom argues that these are political speech, but don't see
> how that stops them from also being source code, or where Tom himself
> indicates that he believes this.

In practical terms, the clause may have dual citizenship. 

However, in his writing, he clearly defines them not as code, but as
political statements. 

> | > 3) Therefore the GPL is not DFSG-free
> | 
> | You fail here. The GPL, by definition of the DFSG itself, states the GPL
> | is free. This is a priori knowledge; defined prior to the start of
> | argument.
> 
> Indeed, that's my point (4).  But the negation of this follows
> necessarily from points (1) and (2) though.
> 
> | > 4) But the GPL /is/ DFSG-free
> | > 5) Therefore either one of the above points is false, or the DFSG is
> | >    contradictory.
> | 
> | The DFSG are guidlines, not philosophy. Its allowed to be
> | inconsistant.
> 
> Yes and no: the DFSG are certainly guidelines, but I think they're
> also an expression of philosophical ideals.  From 
> http://www.debian.org/devel/join/nm-step3 -

Oh, I disagree with you here. There is certainly philosophy going on
behind the scenes, but at no point does anybody advertise that they
deserve the scrutiny of a philosiphical work. 


>       The applicant is expected to agree with the Debian philosophy
>       as expressed in the Social Contract and the Debian Free
>       Software Guidelines.
> 
> I'm also not sure why you think that guidelines should be allowed to
> be inconsistent.

I don't think that, and I hope that they eventually fix it. 

The difference between the DFSG and Tom's essay is that Tom is clearly
trying to take a philosophical approach. As such, his essay must
rigioursly defend against philosophical attack.

My goal here isn't to prove whether or not Tom is correct. My goal here is
to prove that he hasn't met the measure of achiving proof. i.e. He may be
right, and he may be wrong, but he here he didn't prove that he's right.

-- 
James Blackwell          Please do not send me carbon copies of mailing
Smile more!              list posts. Such mail is unsolicited. Thank you!

GnuPG (ID 06357400) AAE4 8C76 58DA 5902 761D  247A 8A55 DA73 0635 7400