[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Discussion - non-free software removal



Hi,

On Fri, Nov 15, 2002 at 07:31:42PM -0500, Branden Robinson wrote:

> On Fri, Nov 15, 2002 at 06:44:41PM +0100, Emile van Bergen wrote:
> > On Fri, Nov 15, 2002 at 10:05:03AM -0600, John Goerzen wrote:
> > > Then you would have us amend the Social Contract:
> > > 
> > >     4. Our Priorities are Our Users and Free Software 
> > 
> > Not necessarily, it doesn't conflict with my interpretation of our goals
> > in any way; it represents it quite accurately.
> 
> That seems demonstrably false.
> 
> "serving the users is in the end the only goal"
> 
> "even the Holy GNU General Public License (and the DFSG) is just a means
> to an end, which is to provide the user with good software"
> 
> It seems that a more accurate expression of your interpretation of our
> goals would be:
> 
> 4. Our Prorities are Our Users and Good Software

No. "Free" is important, because that's how it seems we define Good
Software. We look not only at the software itself, but at the whole of
software + license + its effects on users and developers.

> or even
> 
> 4. Our Priorities are Our Users

That is only implicit; when we say we want only Free Software -- because
we want users and developers to be free, and knowledge to be shared and
reused -- we ultimately only want that for ourselves and other users of
that software.

Computers serve humans. Not the other way around. Software serves no
purpose in itself.

Even for (other forms of) art, it's only worth creating if it can be
experienced by people, even if just by the artist while creating it.

But I don't think we need to change the wording of the social contract,
because it's current phrasing emphasises that we serve our users with
Free Software.

> > However, in the end, software is there only for the humans that use it.
> > Not the other way around.
> 
> I do not see how the existing Social Contract places software before
> humans.  It is humans that use Free Software.  The freedoms that Free
> Software enables are human freedoms.  The freedoms that non-free
> software inhibits are human freedoms.

Exactly. That's why I'm fine with the Social Contract.

> > I don't think that the Debian Project will automagically provide more
> > free software once we stop distributing a non-free repository alongside
> > Debian. That's a too easily assumed argument in this discussion.
> 
> I haven't seen anyone make that assumption.  Can you cite a message
> where a proponent of the GR has done so?

Sure, one by a certain mr. Robinson, who writes in
http://lists.debian.org/debian-devel/2002/debian-devel-200211/msg00985.html

	"It may be that Debian's continued distribution of non-free
	software is actually hurting our users and the Free Software
	community more than we are helping them. Every day a non-free
	package is "better" than a free alternative (or no alternative
	at all) is a day in which less work is potentially done by
	Debian developers and others on a free alternative to create or
	improve it [...]"

> > The end that /not/ removing non-free serves is that it provides users
> > with software that some Debian Developers and some users consider
> > important and free enough to use and support.
> 
> This is consistent with your understanding of the Social Contract, but
> not consistent with what the Social Contract actually says, as far as I
> can see.  According to the Social Contract, the only "free enough" is
> DFSG-free, and under your understanding we'd be better off if we struck
> ", but we will never make the system depend on an item of non-free
> software." from clause one of the Social Contract, among other edits.
> 
> If it's important enough to our users and free "enough" to use and
> support, why *shouldn't* we let the system depend on it, if it makes
> technological sense to do so?

Because *eventually*, we want to have an all-Free system, and we still
care a lot about the distinction.

Just not enough to stop distributing current not-DFSG-free alternatives
through our channels, as long as there are DDs willing to support them.

There is a balance to strike, you know, no matter how hard you make it
sound like it's a choice between selling your soul to the "use
whatever's good for the job" utilitarians or to repent and swear an oath
on St. iGNUcius.

And I think the current situation strikes the balance rather well.

> > If it's useful enough even for some people to spend energy on, who are
> > we to say we aren't willing to lend our infrastructure to allow other
> > users to benefit from that effort, if we don't have to give up work on
> > truly free software to allow it to continue?
> 
> Why not just put all the non-free software in main?

See above.

> > It's not like developers working on packages in non-free will suddenly
> > stop spending time on packages they use and need themselves and start
> > work on some other packages. 
> 
> Some may, others may not.  Do you presume to know with certainty the
> mind of every Debian maintainer who packages non-free software?

No. I only know myself a little and a few others a little less and
extrapolate from there. And you? 

I'm just not overly optimistic about a sudden conception of new
DFSG-free pine-lookalikes and qmail-workalikes and scilab-compatibles
when *we* *merely* *stop* *maintaining* non-free.

They will be written when people care enough about their restrictiveness
to write them. As said, they don't disappear from the earth when the
Debian Project stops distributing them. That fact in itself will not
provide enough of an "itch" to scratch by developing alternatives. Only
the software and its license can do that. Do you really think otherwise?

> > We're all grown ups, we can make our own decisions which software we'll
> > use.
> 
> Then what is wrong with putting this GR to a vote, and deciding?

Nothing. That's also why I'm trying to contribute my views.

> > Of course we believe that people are better off with Free software,
> > but that only applies in general, not for each and every particular
> > situation.
> 
> Then why don't we allow particular packages into Debian main as part of
> our official distribtion, as long as we can legally distribute them and
> they serve a valuable purpose in a specific situation?  This sort of
> approach appears to have worked well for Red Hat.

As said, I like the distinction we make.

> > I won't even spend a blink of an eye thinking if I should get
> > rid of daemontools or qmail on my system if Debian stops providing it
> > through non-free.
> 
> John's proposed GR would not require that you do so.

No. The point is: it won't give me an itch to start impriving postfix
or inetd either.

I think it's quite reasonable to expect that a DD who uses a certain
piece of non-DFGS-free software (which he'll likely continue to do when
it's removed from the non-free distribution), who doesn't use an
alternative Free package now, leave alone work on it, won't start doing
that when that piece of non-free software is not distributed by the
Debian Project anymore.

Exactly because it doesn't disappear from the earth. If that could be
achieved, it might have had some effect on the development of
alternatives. Do you /really/ believe that such development would be
stimulated enough to have noticeable effect when non-free is dropped?

> > I think that dhe Debian project benefits from non-free because Debian
> > plus non-free is more attractive to people than Debian sans non-free.
> 
> I do not see attractiveness mentioned anywhere in the Social Contract.
> 
> > Of course being attractive is not a goal in itself, but important to
> > realise our goal of serving our users (including ourselves) by giving
> > them Free Software.
> 
> I don't understand; under our Social Contract, if it's Free it doesn't
> *have* to be attractive; we can ship it.  So attractiveness must only
> matter for non-free software.

We can ship it. But Debian is only shipped to and developed by people
who are attracted to it. That this is too obvious to be mentioned in the
Social Contract doesn't make it less true.

> > Because to realise that goal, we need developers, and to get developers,
> > we need enthusiast users, and to get enthusiasts we need newbies, and to
> > get newbies we need to provide enough value as a way to get involved
> > with GNU/Linux that people will consider Debian. The fact that Japanese
> > support for xpdf is available, good speech synthesis, DJB's stuff, all
> > helps in that.
> 
> There are few oversights in the above argument.
> 
> * Dropping these things from our mirror network won't stop them from
>   being available, as long as someone wants to make them available.

Just not through the easily accessible, high quality network of Debian
mirrors.

> * Even the presence of DJB's software appears to be of limited utility
>   to us; even when people were furious with the ISC's handling of the
>   recent bind 8 exploits, I saw several testimonials to the effect that
>   they wouldn't switching to djbdns for technical reasons, quite apart
>   from concerns about licensing.  I therefore wonder just how
>   "attractive" DJB's stuff really is.  Of all the non-free software that
>   Debian distributes, I hear the most complaints about his.

I think because there it hurts a lot. His software performs well, is
well modularized, has a good security track record and he makes his
source freely available to use, modify and distribute. It's painful
because he forbids the last freedom, the freedom to distribute
modifications of the source, including the transformation to object
code.

> * Is growth in and of itself a virtue, whether in number of packages,
>   number of developers, or number of users?  If so, how important is it?

Of course not. But if growth is not important, then stimulating the
development of Free Alternatives cannot be the goal of removing non-free
either.

>   What other considerations should we be subordinating in favor of
>   promoting growth?  Perhaps we should drop our principled stance on
>   Free Software, for as you eloquently point out, it is assuredly
>   hamstringing us in some ways.  There is much useful software that
>   would be more readily accessible to our users if we'd just ship it as
>   part of the OS.

Again, I think the current balance Debian stikes is excellent. No
compromises for 'main', which could perhaps be renamed 'Debian', but
still make the rest available through a cleanly separated but just as
accesible channel. 

Cheers,


Emile.

-- 
E-Advies / Emile van Bergen   |   emile@e-advies.info
tel. +31 (0)70 3906153        |   http://www.e-advies.info

Attachment: pgpMzhFX6cNCM.pgp
Description: PGP signature


Reply to: