Newsgroups: rec.arts.int-fiction
Path: news.duke.edu!newsgate.duke.edu!nntp-out.monmouth.com!newspeer.monmouth.com!news.maxwell.syr.edu!colt.net!newspeer.clara.net!news.clara.net!server3.netnews.ja.net!bath.ac.uk!unknown
From: "Ben A L Jemmett" <bal.jemmett@ukonline.co.uk>
Subject: Re: On OS design
X-Newsreader: Microsoft Outlook Express 6.00.2600.0000
X-Mimeole: Produced By Microsoft MimeOLE V6.00.2600.0000
Sender: unknown@bj1084.resnet.bris.ac.uk (Address not verified)
Organization: Jemmett Glover Software Development
Lines: 46
Message-ID: <GMCE7D.A3u@bath.ac.uk>
X-Msmail-Priority: Normal
References: <20011030221953.17603.00000905@mb-cp.aol.com>    <9rohhh$1ib3o$2@hades.csu.net> <9rpl3t$8m9$1@neptunium.btinternet.com>    <Xns914C73A0E79FBknight37m@209.155.56.100>    <ant011834fc4fNdQ@oakseed.demon.co.uk> <9rs8bl$28v$1@news.panix.com>    <n20ur9.m1s.ln@172.16.100.66> <GM71p2.IAu@bath.ac.uk> <knv5s9.2ih.ln@172.16.100.66>
X-Priority: 3
Date: Mon, 5 Nov 2001 19:30:49 GMT
Xref: news.duke.edu rec.arts.int-fiction:94365

"David Given" <dg@pearl.tao.co.uk> wrote in message
news:knv5s9.2ih.ln@172.16.100.66...
> I haven't done any hardware design. (My university was really into the
> theoretical aspects, which is rather a pity.)

My knowledge of the field is all theoretical at the moment too, but I have
the opportunity to do more with it I think.

> But I've seen the various free
> processors you can download and run on FPGAs, and the whole idea sounds
> fascinating.

*nods*  It's been pushed onto my 'ideas to consider at some point' stack,
certainly.  It'd definitely be cool to have a PCB hooked up to a keyboard
and printer, running Z-code natively...

> There you go! The first ever genuine hardware driver for the Z-machine!
Thanks; saved me some effort...  *grin*  But that raises a point; how to do
I/O on a hardware Z-machine?  I guess it'd need some kind of BIOS to deal
with the I/O opcodes[1], but does that have to map into shared memory or
where?  Perhaps a division between 'user memory' and 'kernel memory' is
needed, so the BIOS lives on area of RAM the Z-machine itself can access,
but the gamefile can't...  Then an I/O call is simply mapped by the CPU to a
call into the kernel space...  Uhm, hmm.  Things to think about.  :)

> >> (And before anybody else points it out... yes, I *am* mad.)
> > At least you don't hack around in the GEM Desktop code; now some of
that's
> > mad...  :)
>
> I *used* to. I'm still on the mailing list, but I haven't done much for a
> while.

Oh, I didn't realise you'd worked on the Desktop; I thought that was pretty
much myself and Ken Mauro (with contributions from John perhaps).  It's been
a while since I've done anything with it too, but then it lives in a 64KB
code segment and it's used all but 100-odd bytes IIRC.

[1] I may be completely off-base here, seeing as I've only read through the
Standard once so may have no idea how this works...
--
Regards,
Ben A L Jemmett.
(http://web.ukonline.co.uk/ben.jemmett/, http://www.deltasoft.com/)


