From: teunis <teunis@computersupportcentre.com>
  To  : ggi-develop@eskimo.com
  Date: Fri, 21 Aug 1998 16:08:20 -0700 (MST)

Re: Can I do that ?

On Fri, 21 Aug 1998, Jon M. Taylor wrote:
> On Fri, 21 Aug 1998, Olivier Galibert wrote:
> 
[clip - clip]
> > The so-called  "KGI drivers"
> > should include in fact two modules, one to be loaded by the kernel and
> > one  by libggi.  
> 
> 	Kernel modules cannot be loaded without root privileges.  If you
> ran part of LibGGI as root daemon, it could maybe do something like this. 
> This starts to look like Linus' "superserver" concept.

If anyone's curious this is the direction my old 3D work was going....  it
just wasn't making any sense otherwise (But I'm looking again).

If anyone -really- wants this I'll code it up :)
[won't take long - I already have most of it working :]
(everything outside of graphics is 100%)

> > This  would give much more  room for optimization and
> > experiments in the userspace/kernel communication implementations.
> 
> 	Yes, it would.  I do not recall ever hearing about a dynamic,
> polymorphic, virtual method based OS.  The closest analogy I can think of
> are the class libraries that virtual machine languages like Java and
> Smalltalk are based on.  Anyone know how much of this stuff can be done
> today with Linux modules? 

Actually... *hrm*...  it's duable.  Not simple, but duable.
Think Linux's FS drivers (I think)

Linux is actually quite close.  It's dynamic, somewhat polymorphic, and
has a bunch of virtual method-systems setup...

OS/2 qualifies...  as does QNX afaik.

G'day, eh? :)
	- Teunis