Re: design goals vs mechanisms

[Marcus Brinkmann]

At Thu, 27 Oct 2005 16:08:36 +0200,
ness <address@hidden> wrote:
> Yes. That is right. I really don't have a problem with redesigning the 
> Hurd's architecture. I see this is necessary. But I don't think whether 
> refining the goals is a good idea.

Fair enough.  Maybe it helps if you say what you think the goals are
(irregardless of who defined them), either by reference or in your own
words.

I am asking because I simply don't have a normative document which
states "The goals of the Hurd are (a), (b) and (c)" and which then
says what (a), (b) and (c) are.  This is not a trick question.  I
really don't know.

Once we have some explanation of what the goals are (or should be), we
can take a closer look at them.  Here is my prediction: In all
likelihood they will _underspecify_ the system.  Ie, they don't tell
us everything we need to know to actually design and implement the
Hurd system.  To fill those gaps, we have to add something.  And I
think that "refinement" is one of the things you can add without
breaking with the spirit of the goals.  (Actually, there is a chance
you didn't want to say "refining" but "redefining".  I would agree
with that.  I would not want to redefine the goals of the Hurd, unless
that is supported by the whole project).

Thanks,
Marcus