Hello Peter!

In message <199612081618.RAA00279@line9.slipneu.tu-freiberg.de>
          Peter Berg <Peter.Berg@ww.tu-freiberg.de> wrote:

> Hard disc based booting is *NOT* only a DOS thing! I know no workstation
> which boots from ROM. Also I didn't meen to replace the ROMboot, I only meant 

Isn't it rather the other way round, they all have boot ROMs, often with
password protection and the like, and CMOS configure options about the
hardware? Don't forget diskless machines which fetch everything over the net,
even the kernel!

> to expand it to be *able* to boot from HD, no need to do this if it's not
> needed (RiscOS)!

As Mark pointed out in a later message, you need rather a lot of RiscOS to
decide what to boot, to load the things necessary from the harddisk and so on.

> May be I'm a purist, but I don't get the sence of let two operating systems
> check the hardware if only one will run. It only costs time.

Come on, that's less than 10s or so.
There are so many things which are also not ideal but where you have a real
advantage if you make them more "beautiful".
As long as you don't have serious disadvantages, what's the problem to let two
operating systems check the hardware. You must also know this German proverb:
"If you sew it twice, it will bear more." :-)

> I did this some time before and it worked. But my understanding of a
> operating system is, that it is able to boot itself without the need of
> another OS.
> However ...

It always needs something to get it booted. Would you call the BIOS in IBM
machines call a part of the operating system? Is the bootstrapping code in a
workstation's bootrom already a part of the OS?
Well, so RiscBSD is booted by RiscOS instead of a bootloader in ROM, what's the
difference? What should the reason be to make a distinction between them, to
call one method better than the other?

> All you said above I've done long ago and removed it again because I often
> need to run windows95 and this is not possible from within RBSD. Anyway, pure 

You'll obviously want to swap it, ie. doing nothing boots RiscBSD, holding CTRL
down while booting gets you into RiscOS.

> Anyway, the RiscPC is not a workstation, therefore there is no need to
> discuss this thread further.

I don't think so. It is an issue that it should be possible to run RiscPCs in
"hostile" environments. This also applies to RiscOS though, it should be
possible to safely avoid it from booting unless a password is entered so the
kids at home or the nosey colleague at work can't muck around with your
machine.
I totally agree with Ben Strawson, ART should alter the RiscPC bootphase so
that you can protect it with a password which can only be overridden by setting
a jumper on the motherboard. This is rather trivial to implement, space in the
CMOS RAM for the password permitting.
But even if they implemented it on the spot, we wouldn't get it before the next
update of RiscOS which is still quite a way down the road.
It would not be necessary to change the further boot sequence as long as it's
not interruptable if password protection is on, ie. you may not do a Shift-
Reset then.
To get RiscBSD booted automatically, you would then directly launch it as
before in some obey file. I'm not sure if you can avoid the interruption of
obey files, but you should by locking the Escape key.
A reset at this stage is no problem as this gets you nowhere else than to
another reboot.
The question rather is how you would get into RiscOS again yourself after
you've tried everything to keep the bad guy from doing it. ;-) Well, you'd have
to set the jumper to reallow a Shift-Reset without password into RiscOS.

I wish you all a calm Christmas and a Happy New Year.

Going to bed now,

Markus