[Top][All Lists]

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

Re: The Hurd: what is it?

From: Bas Wijnen
Subject: Re: The Hurd: what is it?
Date: Wed, 9 Nov 2005 17:02:45 +0100
User-agent: Mutt/1.5.11

On Wed, Nov 09, 2005 at 03:00:15PM +0100, Alfred M. Szmidt wrote:
>    > You should go into politics with the evading answers you are
>    > giving, they have no substance other than: Do as you please, I do
>    > not care.
>    >
>    > Which is immensly useful for setting a direction...
>    As you very well know, the current situation is that Mach is
>    considered unsuitable for the Hurd in a releasable state, L4 used
>    to be the alternative, and we are still discussing if it will be
>    used, and if not, what then.
> No I don't.  Nor is this what I was asking.  Nobody has shown that
> Mach is unsuitable; infact all evidence shows otherwise, it is there,
> it works,

I'm not very well educated with respect to Mach.  I have heard from several
people whose judgement in this matter I trust that the IPC mechanism of mach
is too heavy-weight, that we don't use most of it, and that it copies too much
data around.  All that results in a performance which is considered
unacceptable.  I have no reason to believe those people are wrong.  Therefore
I never really looked at the Hurd on Mach, and put my attention to the Hurd on
L4 instead.  If you have good arguments to back up that performance will be
acceptable on Mach, please share them with us.  Until then, I'll continue to
more or less ignore it.

> and it can be improved this `new Hurd' thing cannot.

It can be improved, and it is likely to be even useful, depending on what you
improve, for the "new Hurd".  Of course the "new Hurd" cannot be improved, it
doesn't exist yet.  It can be built though.  But not yet with code.  That will
come later.

> And once again, that is not what I was asking!

No, you weren't asking anything, you were just blaming Marcus for being wise.
What surprised me is that you don't seem to blame Thomas, although he had
exactly the same response in his own words.

>    That means the project is in a very clear state at the moment,
>    which is defining the design principles we want, or if we don't
>    want to work from principles, defining the design goals.  No useful
>    code can be written for the "new" Hurd at this moment.
> So the goal is to sit on our asses waiting for the "new" Hurd to be
> finished?

No, the state is that we *define* the principles and goals.  That's a very
active thing, unlike waiting.  We have to think about it and decide what we

> Marcus made a promise to Jeff Bailey that the Hurd on L4 would be runnable
> last year.

Marcus told you this day that he doesn't like it if you are claiming that he
said things, because you seem to have a habit of misunderstanding him.  Since
this promise was appearantly not made by you, and not to you, I'm not going to
take your word on it.  And since it doesn't have much to do with the
discussion, I don't think Marcus needs to react to it either.

>    As you yourself have stated in the past, we're a bunch of
>    volunteers, nobody tells us what to work on, we choose that
>    ourselves.  That's why the new microkernel needs broad support:
>    Otherwise noone is going to actually build the new system.
> Once again you are totally missing the point, nobody is dictating
> _what_ work you should work on, on you can work on USB support if you
> want, you can sit and improve ext2fs.  But you are working on
> _the_Hurd_ in all cases, and I asked what the Hurd was.

And the answer is "a system which has a running prototype and which is in the
design phase otherwise".  That means if you don't want to design, and only
code, you need to hack on the prototype.  You may not like it, but I don't see
how it is unclear.

> As a volunteer you can decide not to work on the Hurd, and work on something
> else.  Is it a collection of kernels for the GNU system all incompatible or
> is it a single kernel for the GNU system?  If the later, what micro-kernel
> is being used?

That's the next step.  First we have to decide what features we want to
support (or not), and what general principles the design should follow.

> No other project out there has this lack of direction at all.

I haven't seen as much direction in the project since I joined.  We're
defining what we want, then we choose a microkernel, then we implement it.  At
that moment we may loose the direction a bit, since we have to decide what to
implement first, but we'll see about that then.  At this moment, it is very
clear where we are going.  What's unclear is what the result will look like.


I encourage people to send encrypted e-mail (see http://www.gnupg.org).
If you have problems reading my e-mail, use a better reader.
Please send the central message of e-mails as plain text
   in the message body, not as HTML and definitely not as MS Word.
Please do not use the MS Word format for attachments either.
For more information, see

Attachment: signature.asc
Description: Digital signature

reply via email to

[Prev in Thread] Current Thread [Next in Thread]