On Tue, May 21, 2002 at 09:05:32PM +0200, Niklas H?glund wrote:
> Marcus Brinkmann <[EMAIL PROTECTED]> wrote:
> 
> > On Mon, May 20, 2002 at 10:20:27AM -0700, Thomas Bushnell, BSG wrote:
> >   Someone wrote:
> > > > Have you thought about writing a translator for /dev itself, rather
> > > > than using a script to populate a directory? devfs the Hurd way.
> > > 
> > > Certainly we have.  I think this is ultimately the Right Thing, in
> > > fact.
> > 
> > I've just wondered myself about it, but I am not sure how this would be
> > done best.
> 
> Something I would like is complete autodetection of hardware.  It'd be
> cool if a /dev translator could read /usr/share/discover/pci.lst  (or
> whatever) and add subnodes with translators for those devices.
> 
> Alternatively, some tool like discover could be ported to the Hurd to do
> this to a /dev filesystem.

We should have user-space device drivers which live in /dev. Then a
lot of things are possible you probably aren't even thinking of.

> In the long run I'd like to be able to take a bootable CD with the Hurd on
> it, boot it in any modern i386 computer and get a complete configured
> system with a configured X server, mouse, keyboard, scanner, ...

<AOL>Me too!</AOL>

It's just that a lot of work needs to be done for this.

> Then, with some networked filesystem, we've got a truly mobile office :)

Yes, distributed computing is cool. :)

Jeroen Dekkers
-- 
Jabber ID: [EMAIL PROTECTED]  IRC ID: jeroen@openprojects
GNU supporter - http://www.gnu.org

Attachment: msg04758/pgp00000.pgp
Description: PGP signature

Reply via email to