Re: GNU/Hurd Oprating system roadmap inquery.

[Garreau\, Alexandre]

Since Linux exists, the GNU Project declared it as /one of its official
kernels/. Because the aim of the GNU Project is to provide a free
operating system, not a 100% GNU operating system. Linux was the best
way to do it the quickest. Now GNU and the FSF, still in the aim of
freeing users, prioritize projects that free other aspects that are a
lot used, can’t be already freed otherwise and hence make users use
proprietary software and harm their freedom: BIOS (material
initialization in facts), flash player, Skype, etc.

The Hurd is still an “active” (it is, but not enough :/) project, but
since it doesn’t *directely* free users (in the “free” meaning of “free
software”: that is, a right on your software, not the empowerment or the
abilities it gives to you, that’s secondary, it’s the “icing of the
cake” as said rms: the most important is to free what’s not free). So
even the kernel being a important part of operating system in general
(though not so much, I think the most important part is the language —as
for human beings actually x)— and its interpreters, editors and
compilers, and these were already at the core of GNU since its
beginning, and yet GNU has a looot of projects doing a looot of
impressing work here, look at GNU Epsilon for instance), it’s not a so
much important piece of the GNU Project, because its aim is more to make
*a* *free* operating system than *the* GNU operating system.

Yet the Hurd has really great potential, releasing an
“official/reference GNU distribution of GNU” would be useful on the axes
of communication/marketing and could allow us to do a lot of interesting
things without having to stay “just upstream” (there would be less
distance between users and developers).

But the Hurd is still at development level, and the current try (there
were —and several still exists— others: GSRC, swbis, stow, etc.) of GNU
distribution of GNU is GNU Guix, and even if it’s today bootable trough
DMD and Guile, it’s not yet ready to the common user (and more generally
the Guix interface is currently cli, hence enough complex to require
advanced knowledge of its philosophy), so declaring it today as an
official/reference distribution of GNU could be misleading and cause
trouble to users. And yet as said rms here not long ago, risking to say
“the GNU Operating System” speaking about only one distribution would be
misunderstood by the common user who could ignore other distributions
(which could be more likely to fit its need, and give him a better image
of Free Software).

I heard, and the info got confirmed, that it is harder to make things
possible within the Hurd architecture using POSIX (which we need to
implement to meet project objectives) than to make them possible
differently in a Hurd-specific way, so of course it takes some times
(trying to (a) make a revolutionary “kernel” architecture, (b) to be
backward compatible, that’s not that easy), but some of the most
important things works (graphical interface, SATA were /relatively/
recently implemented by youpi too), and others are currently in working
progress (x86_64… and we need USB), also there’s work to make “glue
code” that could allow the Hurd to directly load Linux kernel and
automatically supports everything that Linux supports. The biggest
problem as far as I know, is that most people are contributing really
sparsely, and that the used people are only between 2 and 0, it depends
when. The Hurd is really not anymore an actively developped project, not
as much as other famous projects as Guix or GNUnet.

signature.asc
Description: PGP signature