Re: Hurd queries
- To: Ognyan Kulev <ogi@fmi.uni-sofia.bg>
- Cc: debian-hurd@lists.debian.org
- Subject: Re: Hurd queries
- From: Bob Ham <bob@ham.org>
- Date: Tue, 3 Apr 2001 16:33:01 +0100
- Message-id: <20010403163301.I775@insanity>
- In-reply-to: <20010403153701.A775@insanity>; from bob@ham.org on Tue, Apr 03, 2001 at 15:37:01 +0100
- References: <20010403153701.A775@insanity>
On Tue, 03 Apr 2001 07:32:00 Ognyan Kulev wrote:
> On Mon, Apr 02, 2001 at 08:03:15PM +0100, Bob Ham wrote:
> > Hi all
> >
> > I've a few quesitons about the Hurd that I was hoping could have some
> > light shed on them: firstly, has any thought been put into how sound
> > will be handled? If not, does anybody know if it would be possible to
> > port OSS/Lite or ALSA to the Hurd?
>
> AFAIK There isn't any code for sound yet. If you have desire you can write
> a sound layer and send patches for GNUMach and/or Hurd to bug-hurd@gnu.org.
Quite obviously there is no sound code. This is not what I was asking.
What I was asking is whether or not there had been any discussion on a
sound layer and whether any developers had given any thought to how the
Hurd should handle a sound layer. It looks as though this is not the
case.
> > Is there going to be a Hurd TCP/IP stack developed? I know pfinet was
> > carved out of Linux, and there's a 'newpfinet' directory in the hurd CVS
> > repository (even though it's empty.)
>
> pfinet is TCP/IP stack, isn't it?
Yes. However, it was developed for Linux and not the Hurd (or am I
gravely mistaken?) This means that (amongst other things) what looks
like all of linux/include is included with the Hurd sources and so I'd
imagine there will be code that enables the Linux shaped block to fit
into a Hurd shaped hole. Correct me if I'm wrong, but I'd imagine this
is not a Good Thing.
> But it is tightly integrated with
> ethernet layer, e.g. there is no PPP support yet. I don't know what the
> plans for newpfinet are.
>
> > Has anyone looked at GGI's KGI kernel drivers (ie, what fb came from)? I
> > would imagine that the Mach environment would be ideal to host graphics
> > drivers (as opposed to a monolithic kernel.) Or, are people happy with
> > X?
>
> GGI was mentioned in mailing lists a time ago with the same in mind: as an
> alternative of X. Again, if you like to implement it in GNUMach/Hurd you
> are welcome. BTW X is so popular! But I agree that it's so heavy too. And
> why GNUMach as microkernel to be ideal for hosting graphics drivers?
See Niels Möller's comments.
Also, having a graphics driver is not the same as replacing X. You could
run an X server that uses a Hurd graphics translator. This is kind of
thing KGI/fb with the fb X server does. However, there is also the
Berlin project whose aim *is* to replace X. Berlin seems to me to be a
better choice for a GUI for GNU than X. I say this as everybody knows X
is a *big* beast with lots of issues, both technical and otherwise (X
license stuff.) Combined with the suitablity of the Hurd for a KGI style
graphics system and Berlin's use of GGI (and hence KGI), this seems to
me to be a very powerful arguement.
Bob
--
Bob Ham: bob@ham.org http://pkl.net/~node/
IRC: 'node' on irc.openprojects.net: #slashdot ICQ: 4396425 'node'
"The GNU philosophy is about freedom. To be free one must have
personal power. Personal power is an individual thing, difficult to
obtain and quick to perish." --Krisno Pryosusilo
Reply to: