Ok, I didn't understand most of whats being said here, other than 'fit it
all in an QL' and 'ARM', the first sounds exciting the second would have a
certain person spinning in their grave if 1. they were in one and 2. if it
was implemented :-)
Lee Privett
-------------------------------------------------------------
Sent from my Laptop running XP
but emulating the QL using QPC2
----- Original Message -----
From: "Plastic" plasticu...@gmail.com>
The reason I said ARM is because it's readily available extremely cheaply,
easy to design with, and there's a plethora of available boards already.
As
I understand it, doing something that will run QDOS/SMSQ/E natively would
require either a Motorola 68k family, or certain older Coldfire
processors.
The Motorola route (like the Q40/Q60) is nice'n'all, but doesn't fit the
power envelope I was thinking of, would be hard to embed, and I personally
don't think it has a great future. The high end Coldfire route puts
impositions on the OS to change to conform to changes in instructions and
registers that would be so burdensome nobody will invest the energy.
ARM cores are everywhere, run embedded Linux universally, and can emulate
a
680X0 faster than a 680X0 can run, natively. Current generation multi-core
ARM chips are rivaling Intel performance in a 2W power budget. The only
thing "hard" there is accepting a QL membrane.
Following the ARM route, we can easily obtain ready made boards for around
$100 (70 ukp) complete, or design our own (where are you, Nasta?) and
build
them for around $150 (100 ukp). Separately, we could have our own linux
distribution that strips out EVERYTHING except the ability to run the QL
emulation. That would make it incredibly tiny and fast-loading with a
small
memory footprint and would give us the entire sensation of using a
hardware
QL. A simple utility within the QL could be used to configure how it is
configured, what the emulated hardware profile would be, etc.
Additionally, going this route allows us to expand to have a variety of
different hardware projects that lets us market QDOS/SMSQ/E etc into
various
embedded markets by giving it good data processing/logging and IO
capabilities.
In the longer term future I could see people becoming enamored with the
beauty of ARM assembly when compared against 68k assembly. Given we have
the
source, it's even possible in the future that SMSQ/E could be prepared in
an
ARM native version, even though I know that would not be trivial to do.
QDOS and its children regrettably have no "marketable" future outside this
dwindling community on 68k hardware.
I see the Q60 as the high point of QL native hardware design, but know
such
a venture with the same architecture doesn't get us any further down the
road if it was repeated today. The Q60 is excellent, expensive, and hard
to
make now due to availability of parts. If you were looking to start this
project now, you would probably choose a different hardware base, yes?
My personal opinion. I don't expect people to see it as I do. I'm not
especially fervent about it, I'd just like to see us able to put the QL
operating systems into a more marketable and flexible situation. It's an
amazing opportunity to sell into education, robotics, small industry, and
help the userbase have fatser, cheaper, more flexible kit - that can still
fit in the original case, for spiritual reasons ;)
Dave
_______________________________________________
QL-Users Mailing List
http://www.q-v-d.demon.co.uk/smsqe.htm
_______________________________________________
QL-Users Mailing List
http://www.q-v-d.demon.co.uk/smsqe.htm
