Re: Why licenses don't need to be free (was: Re: Why licenses *are* free)
Hi,
>>"Marcus" == Marcus Brinkmann <Marcus.Brinkmann@ruhr-uni-bochum.de> writes:
Marcus> Base-files contains the GPL for the whole distribution for
Marcus> practical reasons. It can't be removed from main without
Marcus> breaking law.
Rubbish. It can just as easily be moved to verbatim, and no
law is broken.
>> We do not ship the GPL bundled with software. The licenses
>> /usr/doc/copyright/{Artistic,BSD,GPL,LGPL} are shipped sepatrately,
>> in one *ONE* package, namely, base-files (the package is in no way
>> under all those licenses).
Marcus> base-files is essential. This is only for practical
Marcus> reasons. The spirit is there, and we ship the GPL iqn the
Marcus> Debian MAIN distribution because this is required by
Marcus> copyright law.
Nope. All that copyright law requires os that the copyright be
part of Debian. Not that it be included in a particular section of
the distribution (if the copyright law talks about Debian sections, I
shall be intereseted in knwoing about it)
>> We do not ship licenses bundled in with software. Repeat after
>> me: We do not ship licenses bundled in with software.
Marcus> We do. The spirit is there. The difference is only a technical.
Marcus> So, we have to ship the license with the software in the main
Marcus> distribution.
That des not follow. It should be in debian, yes, but not
necesarily in main.
>>
>> Therefore, this statement is also incorrect. We can too put
>> the GPL in verbatim.
Marcus> But we can't remove it from main, and this is my point you
Marcus> choose to ignore.
We can too remove it from main and put it in verbatim. You
choose to ignore this fact too.
Marcus> We don't have a stand-alone GPL package, so this whole
Marcus> discussion is non-sense.
>>
>> Yes, and no. The apckage also contqains other files, including
>> other licenses. But is does not contain any executable that I can see
>> offhand.
Marcus> So what? Removing the base-files package from main would
Marcus> still be illegal.
Rubbish. Removing it from the distribution would be
illegal. The sections have no legal meaning.
>> I think this discussion is not what is non-sense; the
>> arguments you have presented so far may qualify ;-)
Marcus> Is it your only intention to be clever or are you actually
Marcus> trying to be productive in this discussion?
I try pretty hard.
Marcus> You are speaking about stand-alone licenses.
>> Umm, aren't all licenses stand alone? This makes no sense.
Marcus> No they are not. They apply to the work they copyright.
The GPL, by itself does not copyright anything. Various people
have appled the GPL to their code. The realtionship goes the other
way. The GPL is stand alone. The software is not.
Marcus> We have to consider solely licenses that are applied to
Marcus> software and data entities.
>>
>> Yes, true. Like the GPL, artistic, BSD, etc, which we ship
>> stand alone.
Marcus> No, we do not. Stand alone means we could remove them. We
Marcus> can't remove them without being illegal.
Stand alone means we can ship the GPL alone. With nothing else
on the gloppy, I can ship the GPL. It stands alone.
Now, the software is what we can't ship, so the software is
not stand alone.
This is a critical distinction.
>> No more so than standards. The RFC on SMTO doies not itself
>> implement an smtp service, or indeed, talk about internals of a smtp
>> server. It is a meta docuent that talks about the interface.
Marcus> We can choose not to ship a standard, but we can only choose
Marcus> to not ship a copyright if we ship nothing else, too. The
Marcus> copyright is the only thing that grants us redistribution.
Correct. The copyright, however, is stand alone.
>> I see little difference between a standard and a license.
Marcus> This is your problem. Sorry, I have tried to make it
Marcus> understandable for you. If you still can't see the
Marcus> difference, I can't help you.
So, if you can't prove something, the instructor is at fault?
Sorry. If you can't stand behind your arguments, it is *YOUR*
problem. You put up 3 reasons why licenses are diffrent. One was
debatable, two others were plain wrong (as yuou admitted
yourself). And now you have the gall to say you proced it? Since when
do invalid points prove an argument?
Marcus> Is this bad? No. Has this any influence on our work? No. Does
Marcus> it make sense to create an extra section for license
Marcus> documents?
>>
>> I love this. I can now repeat all your arguments about
>> standards here.
Marcus> I take this that you are only discussing for the sake of
Marcus> discussion. This makes sense, as your comments are failing to
Marcus> be reasonable and are deliberately missing the point.
Here goes. These are essentially your arguments. Lets see how
you like them.
There is need for free licenses, as teh NPL and the AbiPL
demonstrate. There is some discussion on whether the NPL is fully
free. If they had been able to derive from the GPL, with a few
additional clauses, this may have been less of an issue.
My fear is that people shall be satisfied with less than free
licenses, especially with the bad example the GPL has set.
The GPL doesn't need to be under copyleft, just a sensible
copyright. IMHO, the GPL should simply say 'Anyone may freely
distribute verbatim copies of this document, as well as derived works,
as long as any derived license clearly indicates its heritage, and
also indicates that it is unaffiliated with the GPL'. Or something
like that.
If you allow modifications or if you use your right to the full
extend. The following has to be kept in mind: ----- With the
name-changing-clause, there will EVER be only ONE version of a
license, the official one. All derivations from it are renamed and
therefore there is no additional confusion created.
Ask yourself, 'Might I, or might one of our users, want to
create a derived work?'
One of the advantages of main, IMHO, is that I know I can create a
dervied work from anything therein, without carefully reading the
license. I would not like to give up that right too easily...
If Debian takes a stand on this, and stands for Free Content in
general (which is really Free Software and all that goes with it), we
could lead this movement. But if we accept the assumption that
licenses, or novels and other artwork need not be free, we may set it
back considerably.
If licenses can't be modified, how can they be improved? I think
there is gain in allowing licenses to be modified. Modified
licenses must be distributed with a prominent notice that this is not
the original license and that the original license may be obtained
from wherever.
But isn't innovation important? If I come up with a new modified
license, and prominently plaster big warnings all over it that this
isn't the original license, why shouldn't I be allowed to distribute
it? Why shouldn't I be allowed to distribute patches so that programs
follow this new license? What if my idea is a good one and the
licenses author see it and incorporates it into the next license?
Is innovation of licenses only allowed to come from the specified
license authors temself?
I don't see why it is *necessarily* a problem if
annotated clearly, and if the derivation does not pose as the
original in any way.
Non free licenses can't be translated. That does hurt the
non-english-speaking free-software community. One needs to know what
one can do with software, but to a non-english speaker a license
written in english is like no license at all. If its author doesn't
allow translations, someone else has to write a new license from scratch.
If everybody choose the "no-translation" terms that means the community
needs different licenses for english, french, german, spanish, italian,
japanese, chinese, ...
Trust is one thing, honesty another. Silly changes to licenses will be
followed by nobody, especially not by Debian.
Taking precautions is one thing, making improvements impossible in the first
place another.
Including anything that is non-DFSG in main, means that people have to start
checking licences, before playing with the source --- a Bad Thing IMHO.
distribution is a question of the copyright: free or non-free
I'm always wondering why people are *afraid* that such a thing could
happen. If the license writers do their job well, the free software
community honours it and follows. And if the license requires a name
change, people who distribute different versions under the same name are
already violating the copyright (so the stricter license wouldn't stop them,
too). I'm always wondering that people are afraid that a license could
be subverted, but nobody is afraid that somebody releases a gcc that adds
back-doors to all executables compiled with it. It is clear to me that
people would only follow a derived license if it is significantly better,
and then there is a problem with the license anyway.
(Actually, I *still* believe that there is a strong case for suggesting
that licenses should allow derived works, with an appropriate
name-change).
Maybe it helps to draw a parallel with software patents? Many proponents
of free software, myself included, feel that their shouldn't be any kind
of IP over 'ideas' or 'algorithms'. I find that, for me, *most*
license documents, technical or not, falls into this category.
let's not talking about compromises (and it is a compromise) before
all facts are on the table.
Great option. Imagine the free software would follow the same criterion. "If
you want to publish a variant C compiler, you can always rewrite gcc".
The whole point about the freeness is that we don't have to rewrite
everything if we want to reuse good parts of other things.
A license will ever be a license, regardless if there are also a few
renamed and changed derivations. If ever one of the derivated works becomes
more popular and known than the license itself, there must be something
wrong with the original license.
In the free community the people are leading who do the work. If the
license commitee does its work well, nobody will feel the need to derive
from the license.
In general, we can expect license documents to be very useful in this
way. A policy that puts them off limits for modification is going to
cause a major problem.
I have a suggestion for how to achieve both goals. This is to permit
modification of the document with the requirement that certain
non-technical sections (the ones that explain that it is a license)
are *deleted* from any modified version.
I think we should add something to the social contract which
encourages 'free' licenses, and explains our rationale for including
non-free licenses, while indicating that we very much prefer free
ones (RMS's point about GNU C went right home for me.)
Marcus> There are now many ways for authors to do this:
Marcus> a) Require a name change.
Marcus> b) Reuqire clear marks where the document has been changed.
Marcus> c) Require original source with diff.
Marcus> d) Require that a non-technical part is only shipped with the verbatim
Marcus> license (see below).
Works for me.
Marcus> If this mail sounds like I'm pissed off by your behaviour in this
Marcus> discussion, take it for granted.
Your being pissed of is your problem too.
manoj
--
"Life's a bitch, and life's got lots of sisters." Ross Presser
Manoj Srivastava <srivasta@acm.org> <http://www.datasync.com/%7Esrivasta/>
Key C7261095 fingerprint = CB D9 F4 12 68 07 E4 05 CC 2D 27 12 1D F5 E8 6E
Reply to: