about GNU Hurd

[Xavier Maillard-5]

Hi,

       Up to date, quality documentation and
       information on the internet is sparse - lots of websites are out of date,
       and now unmaintained.

   Which are web sites you have in mind?  The site for the Hurd is
   http://www.gnu.org/software/hurd; that is where _all_ the information
   should be.

   Would you like to work on updating that site?

I would like to contribute to add a french part. Do you think it
is "needed" or not ?

        Xavier
--
http://www.gnu.org
http://www.april.org
http://www.lolica.org

[Xavier Maillard-5]

   On Mon, Jul 30, 2007 at 08:13:59PM -0400, Christopher Parker wrote:
   > Although I don't think I'd be able to contribute as much, I'd be more
   > than willing to contribute toward a Hurd fund.

   Before jumping the gun, you might want to see how this year's
   Hurd-related Google Summer of Code project works out in the end.  

I agree.

        Xavier
--
http://www.gnu.org
http://www.april.org
http://www.lolica.org

[Xavier Maillard-5]

[Catching up sorry for the long delay]

   In my opinion, what the Hurd needs right now, and first of all, is
   a maintainer which applies patches and, once that works, some sort of
   advertising to attract new hackers.

Alas, I must agree with you. I also tried several times to "enter
the dance" of the Hurd hacking but I never had a chance to see
whether my contributions could be of any help. So I tried to
contact the french Hurd team and here it also seems dead. There
is no activity at all (except conferences and stuff like that).

Given the fact I do not like working for /dev/null, I stopped to
mirror anything. By the way, it would be really cool and indeed,
helps a lot, to dispose of a *real* and *alive* web page so that
any user or interested people could see what's the current status
of the GNU. Currently, there is absolutely no (easy) way to
access to this information. I have the impression that all Hurd
hackers are just « muted ». At last, every time I am looking for
information and status, I feel like I always get the same
answers, years after years I feel Hurd is worst.

I am feeling sad that nobody can take the leadership on a so cool
project, I am also feeling sad to see that the (very) tiny Hurd
community can't discuss the project; there are pro-GNU and
pro-Debian.

Why the FSF or the GNU project does not designate a project
leader ? Why is there no visibility on this project for any
external eye ? What about an official *up to date* website (with
all needed informations to attract new hackers) ? ...

        Xavier
--
http://www.gnu.org
http://www.april.org
http://www.lolica.org

[arnuld-2]

> On 8/28/07, Xavier Maillard <xma@...> wrote:

> Given the fact I do not like working for /dev/null, I stopped to
> mirror anything. By the way, it would be really cool and indeed,
> helps a lot, to dispose of a *real* and *alive* web page so that
> any user or interested people could see what's the current status
> of the GNU. Currently, there is absolutely no (easy) way to
> access to this information.

Thomas Schwinge just created a new Hurd-wiki. I think it will help.

> Why the FSF or the GNU project does not designate a project
> leader ? Why is there no visibility on this project for any
> external eye ? What about an official *up to date* website (with
> all needed informations to attract new hackers) ? ...

I also think so. to me it seems like Hurd not important to GNU
Project, so i stopped reading Hurd User Guide and playing with Hurd on
Qemu. it is pretty much experience of mine that Hurd has
organizational problems and till then they are solved it will not make
any difference to work on technical side. A decade has passed and Hurd
is still an experimental infant :( without any direction where to go
and where to not



--
http://arnuld.blogspot.com/

[arnuld-2]

> On 8/28/07, Xavier Maillard <xma@...> wrote:

> I do not agree Richard. Having the Hurd is essential in that a)
> it would be a GPLv3 project thus protecting user freedom and
> avoiding tivoization and b) would be a GNU (project and kernel)
> and thus would give some guaranties about project quality and
> future.
>


Xavier, this exactly is the same thing i discusses last month, did not
remember i discussed it here or at Debian Hurd mailing list. RMS has
no interest in Hurd.He got stuck with GPLv2 Linux kernel. I think he
has forgotten the fact that there is no GNU OS, most users do not know
about GNU, everywhere is Linux, even saying it GNU/Linux will not
solve any problem because then you deal with effect, not with the
cause. I have to use word Linux to make people understand i am talking
about Linux Operating System, they don't know GNU. majority of users
don't give a damn about Freedom. i talked with many of the users and i
have not found any single user yet who says i will use Free softwares.
 they use Skype and RealPlayer on Linux, what is GNU ?

whoever is going to change this phenomenon  will have to deal with the
cause, not effect. To handle the cause we have only one way, The GNU
OS a.k.a. Hurd and at the pace FSF is developing Hurd, it will take
another 2 decades to get out the 0.1 version and that will be 0.1
version, just that not much. Linus Torvalds does not like GPL
(specifically version 3) and that helps a lot to crush the Free
Software Community using M$-Novell deals and Sun boasting Open Source
Java with no modification and no distribution clauses of JRE
distribution license. We do not have much money to deal with
corporates but we 2 powerful tools to take it down:

1.) Free Software Hackers
2.) Copyleft License

and we need to exploit the 2 to fullest to achieve what we want, a
Free OS, the GNU OS and that will deal with cause not effects, so we
know where to spend our time and resources


> Linux is ok as it is a free software but if they stick with GPLv2
> only, it is a no-no for long term and mainly depends on the "bon
> vouloir" of one person (or a small group of people). GNU's target
> is mainly philosophic and target freedom, linux does not (it just
> has to just work). Having something working is one thing, but
> having something working plus the freedom is much more better.

to be true, we do not have any OS that supports and spreads Free
Software philosophy, that is the Harsh truth-- we do not have Libre
OS, not Linux, i have seen lots of Linux magazines  filled with
proprietary softwares advertisements amd Linus himself supports the
use of proprietary softwares.

RMS has to realize the this fact. He is a good guy and he is the one
who made Linux possible but being a good guy and being able to
understand some requirements are two very different things and that is
why we hear form people Linus Torvalds created Linux Operating System,
what's GNU ?

we really need  an OS that supports Free Software Philosophy ...
I,myself, need it. thanks for spending so much precious time of yours
in reading my opinions :)



--
http://arnuld.blogspot.com/

[arnuld-2]

> On 8/28/07, arnuld <geek.arnuld@...> wrote:

> I also think so. to me it seems like Hurd not important to GNU
> Project, so i stopped reading Hurd User Guide and playing with Hurd on
> Qemu. it is pretty much experience of mine that Hurd has
> organizational problems and till then they are solved it will not make
> any difference to work on technical side. A decade has passed and Hurd
> is still an experimental infant :( without any direction where to go
> and where to not

I correct, it has been 2 decades since 1983 and no GNU OS, no Hurd. If
we can't organize Hurd, if we can't identify and solve the problems
related to GNU OS, then we can abandon the Hurd project at all and put
those precious time and resources into something else. but there is
another way too:  "identify the problems"


--
http://arnuld.blogspot.com/

[Jose E. Marchesi]

   > Why the FSF or the GNU project does not designate a project
   > leader ? Why is there no visibility on this project for any
   > external eye ? What about an official *up to date* website (with
   > all needed informations to attract new hackers) ? ...

   I also think so. to me it seems like Hurd not important to GNU
   Project, so i stopped reading Hurd User Guide and playing with Hurd
   on Qemu. it is pretty much experience of mine that Hurd has
   organizational problems and till then they are solved it will not
   make any difference to work on technical side.

I agree in that the Hurd project has some organizational issues. But,
just my oppinion, those issues maily comes from a lack of capable
hackers, and from a lack of _actual_ work. There are many people
reading the Hurd User Guide, the Hurd Hacking Guide, the OSF Mach
books, etc. But, how many people you know that is able to actually
hack gnumach? From these mach-capable hackers, how many of them are
_actually_ hacking gnumach?

Some years ago there was a spanish Hurd community. We were three
mach-hacking-capable hackers plus several interested hackers using the
system and willing to collaborate. We organized a Hurd Meeting (a
technical one) hoping we could transmit our knowledge about Mach and
Hurd programming to these hackers. Well, the group died a few months
later. The meeting didnt work.

Many more lines complaining about the state of the Hurd project has
been written than lines of code trying to solve it.

Still, there are organizational problems. I think the project need a
refresh. I would appoint Thomas Schwinge as the unique maintainer for
both the Hurd and gnumach. He seems to be the most active hacker doing
real work. He seems to enjoy working in the actual Hurd kernel (not in
HurdNG or something like that). He could make a new website updated
with clear directions and development procedures in
http://hurd.gnu.org. etc.

I would politely ask Brinkman, Bushnell and Neal to pass the baton to
him. They seem to be working in other projects such as HurdNG, that is
not the Hurd anymore since the basic design has changed.

I may be wrong, so just my oppinion :)

[arnuld-2]

> On 8/28/07, jemarch@... <jemarch@...> wrote:
> I agree in that the Hurd project has some organizational issues. But,
> just my oppinion, those issues maily comes from a lack of capable
> hackers, and from a lack of _actual_ work. There are many people
> reading the Hurd User Guide, the Hurd Hacking Guide, the OSF Mach
> books, etc. But, how many people you know that is able to actually
> hack gnumach? From these mach-capable hackers, how many of them are
> _actually_ hacking gnumach?

ver-very few.

well, may be 1 only..

> Some years ago there was a spanish Hurd community. We were three
> mach-hacking-capable hackers plus several interested hackers using the
> system and willing to collaborate. We organized a Hurd Meeting (a
> technical one) hoping we could transmit our knowledge about Mach and
> Hurd programming to these hackers. Well, the group died a few months
> later. The meeting didnt work.

so those mach-capable-hackers refused to work ?


> Many more lines complaining about the state of the Hurd project has
> been written than lines of code trying to solve it.

right

but i can't write code as i don't know C and can't learn it ATM. i am
jobless and penniless, so i am focusing on C++ and OOA-M-&-D as with
that I have lots of opprtunities. If i could get money for C, i could
have started that but that is not possible.

BTW, My personal choice are Common Lisp, Haskell, C and Mercury and if
I ever write a software myself, I will use them


 > Still, there are organizational problems. I think the project need a
> refresh.

right, this is the 1st step we need. we need to identify the problems
and set goals for Hurd. Trust me,when this is done, Hurd will get its
1st version very soon, may be within a year :)


> I would appoint Thomas Schwinge as the unique maintainer for
> both the Hurd and gnumach. He seems to be the most active hacker doing
> real work. He seems to enjoy working in the actual Hurd kernel (not in
> HurdNG or something like that). He could make a new website updated
> with clear directions and development procedures in
> http://hurd.gnu.org. etc.

Thomas Schwinge did a great work of creating a new Hurd-Wiki. He is
doing real-work.


> I would politely ask Brinkman, Bushnell and Neal to pass the baton to
> him. They seem to be working in other projects such as HurdNG, that is
> not the Hurd anymore since the basic design has changed.

I thought that they are no longer working with GNU Porject..

> I may be wrong, so just my oppinion :)

your opinion is right.

--
http://arnuld.blogspot.com/

[Richard Stallman]

    I do not agree Richard. Having the Hurd is essential in that a)
    it would be a GPLv3 project thus protecting user freedom and
    avoiding tivoization and b) would be a GNU (project and kernel)
    and thus would give some guaranties about project quality and
    future.

The thing is, as long as Linux is available and attractive, those that
want to tivoize will be able to do it usinmg Linux as the kernel.
Thus, we cannot prevent tivoization just by protecting the Hurd.

[Richard Stallman]

    I would like to contribute to add a french part. Do you think it
    is "needed" or not ?

If we have pages in French, they should be translations of the pages
in English.  To have pages that are really DIFFERENT in various
languages would be a maintenance nightmare.

[Alfred M. Szmidt]

   I agree in that the Hurd project has some organizational
   issues. But, just my oppinion, those issues maily comes from a lack
   of capable hackers, and from a lack of _actual_ work. There are
   many people reading the Hurd User Guide, the Hurd Hacking Guide,
   the OSF Mach books, etc. But, how many people you know that is able
   to actually hack gnumach? From these mach-capable hackers, how many
   of them are _actually_ hacking gnumach?

One, Samuel.

   Still, there are organizational problems. I think the project need
   a refresh. I would appoint Thomas Schwinge as the unique maintainer
   for both the Hurd and gnumach. He seems to be the most active
   hacker doing real work. He seems to enjoy working in the actual
   Hurd kernel (not in HurdNG or something like that). He could make a
   new website updated with clear directions and development
   procedures in http://hurd.gnu.org. etc.

The most active hacker for a long time has been Samuel, Schwinge has
done very little codewise in total.

[Alfred M. Szmidt]

   > I also think so. to me it seems like Hurd not important to GNU
   > Project, so i stopped reading Hurd User Guide and playing with
   > Hurd on Qemu. it is pretty much experience of mine that Hurd has
   > organizational problems and till then they are solved it will not
   > make any difference to work on technical side. A decade has
   > passed and Hurd is still an experimental infant :( without any
   > direction where to go and where to not

   I correct, it has been 2 decades since 1983 and no GNU OS, [...]

Anyone using GNU/Linux uses the GNU operating system.  I once
suggested that we ditch the Hurd, and use Linux instead since Linux
already provides everything the Hurd does and ever will; nobody caught
onto that idea.

Try to catch the wind, as the song goes...

[Xavier Maillard-5]

      > Why the FSF or the GNU project does not designate a project
      > leader ? Why is there no visibility on this project for any
      > external eye ? What about an official *up to date* website (with
      > all needed informations to attract new hackers) ? ...

      I also think so. to me it seems like Hurd not important to GNU
      Project, so i stopped reading Hurd User Guide and playing with Hurd
      on Qemu. it is pretty much experience of mine that Hurd has
      organizational problems and till then they are solved it will not
      make any difference to work on technical side.

   I agree in that the Hurd project has some organizational issues. But,
   just my oppinion, those issues maily comes from a lack of capable
   hackers, and from a lack of _actual_ work. There are many people
   reading the Hurd User Guide, the Hurd Hacking Guide, the OSF Mach
   books, etc. But, how many people you know that is able to actually
   hack gnumach? From these mach-capable hackers, how many of them are
   _actually_ hacking gnumach?

The lack of knowledge justifies the lack of skilled people. Very
few people just know what the Hurd is and its internals.
Whatsover, as I told it, there is no easy way to just know what
to do and how to do it well. In other words, nobody can easily
get access to valuable knowledge to work on the Hurd. This is
more important than just an organizational problem.

Each time I tried to get interest in the Hurd, things were
changing (l4, gnumach, coyotos, etc.). Due to lack of a « kind
dictator », project lacks directions on what things should be
done. Plus there are big dissentions between project members so
that's really hard to stay on the project.

   Many more lines complaining about the state of the Hurd project has
   been written than lines of code trying to solve it.

I agree with that.

   Still, there are organizational problems. I think the project need a
   refresh. I would appoint Thomas Schwinge as the unique maintainer for
   both the Hurd and gnumach. He seems to be the most active hacker doing
   real work. He seems to enjoy working in the actual Hurd kernel (not in
   HurdNG or something like that). He could make a new website updated
   with clear directions and development procedures in
   http://hurd.gnu.org. etc.

   I would politely ask Brinkman, Bushnell and Neal to pass the baton to
   him. They seem to be working in other projects such as HurdNG, that is
   not the Hurd anymore since the basic design has changed.

Lucky you, there is at least someone that knows the basic design
behind the Hurd ;) What I think must be clearly told is this:
will gnumach survive or not ? There were so many (controversial)
things said about this that I still do not know what the real
status is. When I first began to (try to) code on the Hurd
project, gnumach was closed to be dropped for L4. Then I started
to read on L4 and I saw L4 would have been also dropped for
something else. Hard to stay motivated given the facts that OS
programming skills are hard to acquire and so are informations
about each of these micro-kernels, in the end I felt like a dumb.

So I totally agree we should follow one and only one leader and
just do the work.

   I may be wrong, so just my oppinion :)

This is your opinion and we _must_ respect it :)

        Xavier
--
http://www.gnu.org
http://www.april.org
http://www.lolica.org

[Richard Stallman]

Marcus has not abandoned the Hurd.  He is working on a new
microkernel, which he thinks we need.  I hope that this will turn out
well, but one cannot tell in advance.

Meanwhile, I could give more responsibility to Thomas if he wants more
responsibility.  Thomas, do you want it?

[Xavier Maillard-5]

   Marcus has not abandoned the Hurd.  He is working on a new
   microkernel, which he thinks we need.  I hope that this will turn out
   well, but one cannot tell in advance.

Out of curiosity, what is this new microkernel name ?

        Xavier
--
http://www.gnu.org
http://www.april.org
http://www.lolica.org

[olafBuddenhagen]

Hi,

On Fri, Aug 31, 2007 at 03:00:17AM +0200, Xavier Maillard wrote:

>    Marcus has not abandoned the Hurd.  He is working on a new
>    microkernel, which he thinks we need.  I hope that this will turn
>    out well, but one cannot tell in advance.
>
> Out of curiosity, what is this new microkernel name ?

It doesn't even have got a name yet. Really, don't hold your breath.
Nobody -- including Marcus himself -- knows whether something will come
of it. There are good reasons why he hasn't announced it. It's an
experiment.

If you want to work on the Hurd, work on Hurd/Mach -- which still is the
mainline implementation, and very likely will remain so for years to
come. Hurd/Mach is not going away any time soon, regardless of any
experiments with possible replacment microkernels. And even if one day
that changes, the work on the existing implementation isn't lost. These
experiments are not an excuse not to work on the Hurd as it is.

-antrik-

[Richard Stallman]

    Lucky you, there is at least someone that knows the basic design
    behind the Hurd ;) What I think must be clearly told is this:
    will gnumach survive or not ?

When I discussed this with Marcos a couple of years ago,
he pointed out serious flaws with GNU Mach.  He believed that
replacing it was essential.

The flaws are real.  I recall that at least in part of that discussion
I thought that the flaws were not absolutely fatal, that the Hurd on
GNU Mach could still be useful for some things.  I don't recall what my
conclusion was at the end of the discussion.  But it is clear that our
best hope for making the Hurd really good is Hurd-NG.

[Xavier Maillard-5]

       Lucky you, there is at least someone that knows the basic design
       behind the Hurd ;) What I think must be clearly told is this:
       will gnumach survive or not ?

   When I discussed this with Marcos a couple of years ago,
   he pointed out serious flaws with GNU Mach.  He believed that
   replacing it was essential.

Indeed this is what I did understand too but as time passes,
there is no news on what is *currently* done and *how*.

   The flaws are real.  I recall that at least in part of that discussion
   I thought that the flaws were not absolutely fatal, that the Hurd on
   GNU Mach could still be useful for some things.  I don't recall what my
   conclusion was at the end of the discussion.  But it is clear that our
   best hope for making the Hurd really good is Hurd-NG.

Let see how this is going to be. I hope we will see something out
not that too long.

        Xavier
--
http://www.gnu.org
http://www.april.org
http://www.lolica.org

[Michael Heath]

I think the problem people face is: Why work on something that is already fundamentally obsolete or bad? Why not focus on the reimplementation you know has to come?

And because questions of what that reimplementation is and how people can help with it keep being answered with "I don't knows", "Don't you worry about that", or half answers, people aren't excited to help with anything.

Other projects have a clear road map, clear set of design goals, and a clear structure on who is in charge of what - why doesn't the Hurd and associated projects?

Road maps are important. I meet people every day who are absolutely convinced that the Hurd doesn't even exist, really - that there is no usable implementations of, that it doesn't boot, or do anything useful. Road maps can give people a clear idea of what your projects existing capabilities are, and also tell them whats going to come in the future and when it can be expected.

Just my two cents,
Michael Heath

On 9/2/07, olafBuddenhagen@... <olafBuddenhagen@...> wrote:

Hi,

On Fri, Aug 31, 2007 at 03:00:17AM +0200, Xavier Maillard wrote:
> Out of curiosity, what is this new microkernel name ?

It doesn't even have got a name yet. Really, don't hold your breath.
Nobody -- including Marcus himself -- knows whether something will come
of it. There are good reasons why he hasn't announced it. It's an
experiment.

If you want to work on the Hurd, work on Hurd/Mach -- which still is the
mainline implementation, and very likely will remain so for years to
come. Hurd/Mach is not going away any time soon, regardless of any
experiments with possible replacment microkernels. And even if one day
that changes, the work on the existing implementation isn't lost. These
experiments are not an excuse not to work on the Hurd as it is.

-antrik-

[Jose E. Marchesi]

   I think the problem people face is: Why work on something that is already
   fundamentally obsolete or bad? Why not focus on the reimplementation you
   know has to come?

   And because questions of what that reimplementation is and how people can
   help with it keep being answered with "I don't knows", "Don't you worry
   about that", or half answers, people aren't excited to help with anything.

   Other projects have a clear road map, clear set of design goals, and a clear
   structure on who is in charge of what - why doesn't the Hurd and associated
   projects?

   Road maps are important. I meet people every day who are absolutely
   convinced that the Hurd doesn't even exist, really - that there is no usable
   implementations of, that it doesn't boot, or do anything useful. Road maps
   can give people a clear idea of what your projects existing capabilities
   are, and also tell them whats going to come in the future and when it can be
   expected.

100000% agreed.

If we are sure we need HURD-NG, it is no-sense to work in Hurd on
gnumach anymore. To ask people to work in Hurd on gnumach is then to
ask hackers to work in something that wont be useful.

If we are sure we need HURD-NG, lets call it HURD, stop the
development of the current Hurd on gnumach, write new webpages, make a
new development group able to help Marcus, and replace the current
project. Having people working in the current Hurd project waiting for
rumours about the new microkernel doesnt seems very logic to me.