Why licenses *are* free (was: Re: Why I don't share Manojs fears.
On Fri, Aug 14, 1998 at 10:26:01AM -0500, Manoj Srivastava wrote:
> Hi,
>
> All Right!. This is a good message; something I can sink my
> teeth into.
Thank you. I'll apologize for all fuss I created. This is a difficult
issue, and it was hard for me to get my thoughts sorted out.
> Something to consider: you have gone a ong way towards
> convincing me that the rename-and-distinguish-if-modified clause is a
> good thing if one is creating a standard; but what if the author has
> gone the route of no modifications at all? Is this so egregious that
> we throw it out of Debian?
Well, this depends how much weight you give to the advantages and
disadvantages and is essentially opinion. For me, the consequence is to put
them in non-free (which does not mean that we can't obeye them). What would
this express? This would express that the current situation is in some way
sub-optimal, but this does not mean that we can't live with a slight
dependency on non-free, as we do with pgp.
Many people think there should be a section for immutable documents.
I consider this solution as sub-optimal. Either we are in favour for
standards you can derive from (which name-change-and-whatever), or we are
not. Creating another section is sort of laying the problem aside, trying
to not adress it properly. It can also be seen as a compromise, and if the
majority wants to see it this way, I'll not object, assuming that you can
come up with a good definition what belongs in the new section.
My fear is that people will be satisfied with documents belonging in the
"verbatim" section. For example, if software documentation also is allowed
to belong there. I would like to see a proposed definition of the new section
before I express further considerations.
> Why should exeptions be made for licenses
> (which are in just as great a need to be improved and modified look
> at NPL, AbiPL, and other GPL knockoffs [which in some way do dilute
> the GPL, is only psychologically]) but not for standards?
I'll adress this further below. Just read on.
> What about
> non-technical documents, like Graphic novels or any of the other
> categories?
We haven't talked about them yet. Currently, we don't ship such things, do
we?
> Actually, I am going to make a stand about our Hypocrisy;
> anything that you have said also applies to Licenses. You want to
> throw things like the FHS and others out of main, you have to throw
> out the DFSG, the social contract, and GPL etc out as well
>
> All the arguments about stadards apply to licenses as well.
This seems to be your core argument in the current discussion, but it is
wrong or right, however you want to see it. But what is wrong, is the
conclusion that copyright licenses need to be free.
Reasons why Copyright documents are different from standards.
1) Practical reason
Every piece of software comes with copyright notices. A big deal of Debian
is under the GPL. If we choose we can't ship the GPL, we can't ship those
software. For example, we couldn't ship dpkg :) and all FSF stuff.
2) Legal reason
No copyright can restrict you on what license you choose to put your work
under. This means, whatever license I write, I don't violate a copyright.
This applies mainly to the legal text, of course. So, I can take the legal
text of the GPL (the "terms"), and apply them with whatever changes I want
to my document. This means:
* Deriving a new license from the GPL using the terms of the GPL is already
granted by common law. *
3) Technical reason
The GPL is not a technical document. The only benefit you can gain from
taking parts of the GPL is taking from the legal text, which is already
granted by point 2 above. This leaves the preamble and the footer, both are
very FSF specific and do not contain any legal or technical text, but
personal opinion and meta-text. Does somebody wants to derive from that?
Maybe. Is it essential and useful? Probably not.
Shouldn't we ask RMS to make the license of the GPL more free?
No. Because it alöready is as free as it ever needs to be (see point 2
above). Essentially, if RMS would change the copyright of the GPL text, he
would probably choose something like this:
Fictional GPL copyright:
* You may distribute verbatim and modified versions of the legal terms
below, but you have to remove the preambel and the section "How to
apply..." from the derived work. You have to remove any reference to the
FSF and GNU GPL, too *
This would essentially be a free+(name change, remove non-technical part)
copyright, and it is exactly that what is already granted by law. It is not
necessary to include all permissions already granted by law in the license
text. For example, fair use is also not mentioned in the copyright
documents, but is granted anyway.
We should judge the freeness not by the verbatim words of the copyright, but
by the words of the copyright in context with common law. Using the first,
the GPL may appear non-free, considering the common law reveals that it is
as free as we want.
There is a further argument, I'm not sure people will agree with. However,
for the sake of completeness, I'll name it here:
Copyright documents are very short, compared to standard documents.
Rewriting a copyright from scratch, using quotes of legal texts (see point 2
above), is done fast. Rewriting a standard from scratch can take many
months. However, I think this argument is not needed in this discussion, I
find point 1-3 above already compelling.
About the DFSG, the social contract et al:
Those are neither licenses nor technical documents. They are political
texts, opinions, meta documents or whatever you like to call them. They are
short, they are Debian specific. Using parts of them is already granted by
fair use. I can't see what they have to do with copyrights, or even with
standard documents. Those documents play in different catogories, we haven't
talked about yet.
BTW: If the dfsg is not free, is the Open Source definition in violation?
> >>"Marcus" == Marcus Brinkmann <Marcus.Brinkmann@ruhr-uni-bochum.de> writes:
>
> Marcus> On Thu, Aug 13, 1998 at 11:53:40PM -0500, Manoj Srivastava wrote:
>
> Marcus> please accept that I see it differently from you. You think
> Marcus> that modifying a standard is detrimental to the community, I
> Marcus> don't see it this way.
>
> Firstly: having differing views is a good thing. Secondly, my
> fears are indeed more about dissemination of almost similar standards.
Are there further restrictions you want to allow in main beside the four we
already have? (name change, marking changes, orig source+diff and the
requirement to remove a non-technical part of it are already named. I'm
happy to extend the list). I think a derived work from a standard document
that uses all these restrictions can't be very similar to the real standard
anymore...
> Marcus> BTW: I don't think reuse is a major goal, I think the
> Marcus> possibility to reuse is an important thing. For example the
> Marcus> Docbook Standard: It is a DTD which defines a standard format
> Marcus> to write documents in. It is extensible, but you can also
> Marcus> modify the core DTD, you just have to rename it, and give it
> Marcus> a different pubic identifier. Most people will use the DTD as
> Marcus> is, but if there is the need, you don't have to start from
> Marcus> the beginning.
>
> All right. This may be acceptable.. I will even grant you that
> derived works, as long as the original is not tainted with knock-off
> copies pretending to be the original, are beneficial to the
> community.
Knock-off copies pretending to be the original would be a copyright
violation even of the free licenses+named restrictions. The legislation
can't protect us against criminality, there is the executive and the
jurisdiction for that.
> Marcus> Raul Miller:
> >> It's important to recognize that DFSG already provides a mechanism which
> >> preserves the integrity of a standard: that's the "you must relabel
> >> the document if you change it" type of license.
>
> Not merely relable, but remove all vestiges of the original
> name. But I concede this point. Applies to licenses as well.
Yes, a name change would not be enough in all cases. Luckily, we have more
acceptable restrictions available, are there more we can accept?
> Marcus> Adam Harris:
> >> Well, suppose we want to add an appendix to the FHS. For instance,
> >> the FHS doesn't not talk about icons and pixmaps and where shared
> >> pixmaps should be placed. We should feel we have the power to add
> >> components to the standard; in this case, probably it would mean a
> >> separate *appendix* document. However, I could see cases where we
> >> might feel that for the benefit of the developers, it's easier for
> >> them to look at the FHS, and our extensions (still compliant with
> >> baseline FHS) in the same document. So couldn't we, shouldn't we, be
> >> empowered to retitle the document ("Filesystem Heirarchy Standard,
> >> including Debian extensions"), and add a few additional directories,
> >> in each case where we are adding, mark the addition as Debian-specific
> >> very clearly?
>
> Bad. Really Bad. This is not renaming. This effect is better
> achieved by creating a new documents, and saying that we follow the
> FHS, and we also follow Policy, and that document is part of
> Policy. We should never modify the original and call it Debians
> version of ``Original''; there should be no chance of confusion.
I agree. Deriving from a standard should only be done in cases were it is
useful, for example to write software documentation (example: ANSI C and GNU
GCC), or when we want to keep a "died" standard alive. So, the example above
is not necessarily a good example.
> Marcus> [...]
> >>
> >> IMHO, locally modified versions of a standard are better than nobody
> >> following an insufficient standard at all and everyone making up his own
> >> "standard". Eventually someone will step up and merge all differences into a
> >> single standard again (as it happened with apache, for example. Sorry for
> >> the software analogy).
>
> Following a locally modified standard is worse than no
> standards at all. Yuo modify what the standard says, and you follow
> the modified stnadard, you are harming yourself, and not following
> what evceryone else is, and you are harming the community.
Granted. Especially, you are harming yourself, and this is the reason why I
think many people will stick to unchanged standards even when standards are
free+acceptable restrictions.
Thank you,
Marcus
--
"Rhubarb is no Egyptian god." Debian GNU/Linux finger brinkmd@
Marcus Brinkmann http://www.debian.org master.debian.org
Marcus.Brinkmann@ruhr-uni-bochum.de for public PGP Key
http://homepage.ruhr-uni-bochum.de/Marcus.Brinkmann/ PGP Key ID 36E7CD09
Reply to: