[GNU-linux-libre] Re: any Free BSD variant?

[Yavor Doganov]

Rubén Rodríguez Pérez wrote:
> For power level this is true: the more you do, the more you rule. 

You are confusing the rules which each project applies with the
fundamental freedoms each free software user has.

> So you are not the average user

Maybe I'm not, but this doesn't mean I don't consider other POVs.  I
was helping users all off my life in various ways; I'd wish they
caould learn how help themselves, because it's already possible.
It was possible since day 1.  That's all.

> That's where I think you are wrong with this. Compiling is for
> advanced users,

I think you're spoiled by the same "desease" that I'm spoled of.  Tell
this to the Gentoo/Ututo or *BSD users.  As a matter of fact, I often
have to compile packages manually in gNS.

If it is required to have developer skills to compile Emacs or any
other package, then that package is buggy.

Now, the gentle art of debugging the NSDistributedNotificationCenter
class in GNUstep Base or the GNU ObjC Runtime is something different,
and having such skills would make one a developer, most probably.

> Your knowledge level is way above the average computer user. 

Your opinion is based on a few things I wrote that seem technical,
more or less.  I'm lacking lots of skills I think everyone should
have, and you prematurely "classified" me just because I said
something that sounded a bit technical.

> Yes, a good distro should include the best of the two worlds: 

I agree.  I just wonder if what I said sounded like a rock monastery
for hard-core cavemen and dedicated monks with all the console
goodness by default?  That's not it.  I want the "opportunity" for
such a humble life, which is very difficult with web-based systems
like LP, Savannah or RT.  It is veru easy to achieve that with
Debian's infrastructure, beacuse they *care about such monks*.

> Then you need to explain what do you mean with "pure GNU".

I didn't say "pure GNU", did I?  I don't think there should be an
"official GNU distro", for mostly the same reasons rms thinks this is
not a feasible idea.

> I thought you where talking about just the GNU packages.

Yes and no.  GNU packages + some extra packages form the GNU system.
What triggered this specific debate is my preferrence to the GNU
implementation of certain established non-GNU alternatives.

> > As a GNUer, don't you want to improve the system?

> Of course I do, but the GNU packages are just a tool.

They are much more than "just a tool".  If properly done, the metadata
provided in the package can encourage users to learn about the goals
of the GNU project, and eventually support them, even occassionally or
in little ways.  That still counts for something.

As a mortal user, I've been choosing packages by this criteteria
(mostly), as suprising as it may be.

> The only matter is the philosophical one.

This is the most important one for me, yes.

> So, if a free program exists, then the hackers should not waste
> their time by doing a GNU version, 

You can't say what hackers should or should not do :-)  Entire fairly
rich and complete APIs are being wiped out as we speak because the
designers consider it's worth redesigning, and sometimes this is the
only way to solve certain bugs and deficiencies.

I agree with you that developing free replacements should be the
priority, but that's not you and me are doing, right?  Why?  Because
it's hard, and because personally I'm not capable to contribute in
this area.  I fully agree with you that writing a free replacement of
a non-free program is higher priority than whatever system design
rules we can come up with.  But I can't realistically do it, so I'm
sticking to what I think I can do.

> In most cases, that solidarity is mostly filled with ego too.

I disagree.  What ego are you talking about?  Brian decided that Emacs
should be installed by default in gNS, and he never contributed to
Emacs, AFAICT.

This is not a question of ego, but exactly solidarity.  When I mainain
bar and you maintain foo, and foo depends heavily on bar and bar is an
obscure package with limited real users mostly behind the scenes, I'll
make good effort to cooperate with you, the fellow meaintainer of foo,
because you're in the same camp and at the same time the heaviest
user.

A perfect example of such natural cooperation is how the GNU Build
System is being developed.

> As a distro maintainer, I choose GNU programs if they are equally good
> as some other.

Let's be honest: you've never evaluated the GNU implmentations, you
just accept the alternative that your upstream ships (I don't even
know which you distro you work on, FWIW).  Because, as was pointed out
by Karl, choosing the GNU alternative is a price you can't afford to
pay right now, for obvious reasons.  When your developrs surpass 30,
maybe you could :-)

> True, In this case, replacing a commonly used free app with a not
> so known GNU package is a pointless effort anyway, even it the GNU app
> already exists.

No, it would result in a flood of bug reports about use cases when the
developer of the package knows nothing about.  It would improve the
integration between GNU packages immensely, much like the
error-parsing work done in the early days.  That's a good thing.

> The same applies to compiling too ;)
 
Again, tell this to the Gentoo people :-) They consider it a good
thing to have full control over the compilation process, and I can
only understand them.  (If you don't know a punushment for a Gentoo
user, just let him compile anything on my hardware :-))

Saying that such a system is user-unfriendly is not right: Even in my
(proprietary) country and city many companies support corporations
this way.  So it is clearly something that works.

> What I meant is that you cannot replace those packages with GNU pieces,
> because we don't have them. Ok, you can use the HURD, but we have no X.

Have you used GNU/Hurd?  Hurd is not a replacement of Linux because it
has never worked properly (even now, and many of the Hurd design
super-features were implemented in Linux, one way or another).  X was
never in the list of packages to reimplement/replace, and even Hurd
wouldn't be there if Linux was as old as X.