Re: Let's do some coding :-)

[Marcus Brinkmann]

Alfred,

At Tue, 25 Oct 2005 10:09:39 +0200,
Alfred M Szmidt wrote:
> You obviously don't get it, the core of the GNU system is not the
> Hurd.  The Hurd must tie into the rest ot the core GNU system which
> usually consists of quite alot (coreutils--implements POSIX bits,
> glibc--implements POSIX bits, bash--implements POSIX bits).

I don't know what the Hurd must or must not, but I know two things:
This is not for you alone to decide, and it is premature to say if
what we are doing here will satisfy all requirements the Hurd must
satisfy or not.

In other words, keep your side sweeps at home, they are uncalled for
and we don't take them from you anyway.

What absolutely and totally doesn't work is to try to impose a
thinking prohibition on us just because of some self-proclaimed
restrictions on the Hurd design goals.  We will try to envision the
best system we can build, and _then_ we will see what the consequences
are.  _If_ the consequences will be that it cannot be used for GNU,
_then_ (and only then) we will examine the reasons for this and see if
there are fundamental incompatibilities, or if it is just an
engineering problem.  

Note that I don't expect any problems of this sort.  In fact, the
POSIX compatibility layer part of the design is pretty easy to
understand and there is sufficient precedence, including the Hurd on
Mach, to believe that it can be pulled off.  This is currently not
under discussion, just because it is the easy part, and we are right
now struggeling with the hard, non-POSIX, parts.  If you are only
interested in the POSIX parts, then this is a strong signal for you to
stay out of the way.

If you have troubles understanding the design, or see some problems
implementing it, then you are welcome to ask technical questions about
it.

Thanks,
Marcus