Looking for a hardware challange ?

[Jos Dreesen]

Here are the Lilith schematics in Eagle format....

ftp://jdreesen.dyndns.org/ftp/CAD


If anyone is crazy and/or foolhardy enough to create new PCB's, then I want a set !


Jos

[Chris Burrows]

Post by Jos Dreesen
Here are the Lilith schematics in Eagle format....
ftp://jdreesen.dyndns.org/ftp/CAD
If anyone is crazy and/or foolhardy enough to create new PCB's, then I want a set !

BatchPCB would probably do it for you at $2.50/sq. inch (assuming they are
no more than 2-layer boards , maximum dimension 10" x 15")

http://www.batchpcb.com/

--
Chris Burrows
CFB Software
http://www.cfbsoftware.com/modula2

[Fruttenboel]

Post by Chris Burrows

Post by Jos Dreesen
Here are the Lilith schematics in Eagle format....
ftp://jdreesen.dyndns.org/ftp/CAD
If anyone is crazy and/or foolhardy enough to create new PCB's, then I want a set !

BatchPCB would probably do it for you at $2.50/sq. inch (assuming they are
no more than 2-layer boards , maximum dimension 10" x 15")
http://www.batchpcb.com/

I have extremely good experiences with http://www.pcbcart.com.
unfortunately the AMD bit slice processor will be very hard to
obtain... so the chances of creating a new lilith are minimal.

it would be better to rewrite the software for the pentium I processor
and the pc platform. better and cheaper.

[Chris Burrows]

Post by Fruttenboel
I have extremely good experiences with http://www.pcbcart.com.
unfortunately the AMD bit slice processor will be very hard to
obtain... so the chances of creating a new lilith are minimal.

Not as hard as you might think. There are quite a few listed on eBay stores
for $10 - $20. e.g.

http://stores.ebay.com/ACP-Surplus

I have dealt with them before and can recommend them.

Post by Fruttenboel
it would be better to rewrite the software for the pentium I processor
and the pc platform. better and cheaper.

Nah! What would the challenge be in doing that?

--
Chris Burrows
CFB Software
http://www.cfbsoftware.com/modula2

[Fruttenboel]

Post by Chris Burrows

Post by Fruttenboel
it would be better to rewrite the software for the pentium I processor
and the pc platform. better and cheaper.

Nah! What would the challenge be in doing that?

that's what the real men did with aos: port oberon to the dreaded x86
environment. all other platforms are irrelevant nowadays. x86 is the
borg. resistance is futile. they will all be assimilated.

it will still be a hell of a job to make a stand alone lilith86
system. enough of a challenge for a mortal soul. at least, in the
northern hemisphere it would. :o)

[Jürgen Lerch]

Saluton!

Post by Fruttenboel
that's what the real men did with aos: port oberon to the dreaded x86
environment. all other platforms are irrelevant nowadays. x86 is the
borg. resistance is futile. they will all be assimilated.

Of course, the Borg have been overcome ...

Ad Astra!
JuL

[Christoph Schlegel]

Post by Jürgen Lerch
Saluton!

Post by Fruttenboel
that's what the real men did with aos: port oberon to the dreaded x86
environment. all other platforms are irrelevant nowadays. x86 is the
borg. resistance is futile. they will all be assimilated.

Of course, the Borg have been overcome ...
Ad Astra!
JuL

Yep, I remember someone saying: "Assimilate this" - meaning something
different...

[Christoph Schlegel]

Post by Fruttenboel

Post by Chris Burrows

Post by Fruttenboel
it would be better to rewrite the software for the pentium I processor
and the pc platform. better and cheaper.

Nah! What would the challenge be in doing that?

that's what the real men did with aos: port oberon to the dreaded x86
environment. all other platforms are irrelevant nowadays. x86 is the
borg. resistance is futile. they will all be assimilated.
it will still be a hell of a job to make a stand alone lilith86
system. enough of a challenge for a mortal soul. at least, in the
northern hemisphere it would. :o)

I am quite sure Native Oberon ran on my mighty x86 (486) box long before
aos appeared, so I think they didn't port but developed aos on top of
quite portable code.

[Fruttenboel]

Post by Christoph Schlegel
I am quite sure Native Oberon ran on my mighty x86 (486) box long before
aos appeared, so I think they didn't port but developed aos on top of
quite portable code.

that is definitely true. i used to run NO on a dedicated laptop. it
even ran bluebottle.

but, i got bored of oberon, being OS and compiler in one. oberon is a
fine language but the os is clumsy (compared to the one with the
penguin).

and, this opens an alternative: if one has Oberon0 booting, it should
be possible to rewrite Lilith in Oberon0 and some 386 assembler. just
put the dreaded x86 in flat real mode and you have the finest of
processors one can imagine.

just use the keyboard, mouse, vga, ide, sata, etc of the x86 platform
and concentrate on lilith. get a simple system running, with data and
sources supplied on floppy disk or CDrom for starters. then add ONE
brand/type of NIC's (i propose the RTL8139 chip which is available for
ridiculously low prices) and set up a tcp/ip stack.

it could be done. it might be feasible.

it could even be done with most of us for free. i guess i won't be the
only one here who still has an asus xp55t2p4 mobo lying around, loaded
with a cpu and (for lilith) lots of ram.

[Chris Burrows]

Post by Fruttenboel
and, this opens an alternative: if one has Oberon0 booting, it should
be possible to rewrite Lilith in Oberon0 and some 386 assembler. just
put the dreaded x86 in flat real mode and you have the finest of
processors one can imagine.

Why rewrite it in Oberon - there's plenty of Intel Modula-2 compilers
around. Then again, why bother to do that. The M2M-PC system runs the
original Lilith compiler and Medos on top of an M-Code interpreter which was
written in x86 assembler. Copies can be downloaded from my site:

http://www.cfbsoftware.com/modula2

If the Lilith OS and applications were ported to an Intel Modula-2 compiler
(NOT a trivial task due to the differences of PIM 2/3/4 and ISO M2) it would
certainly run faster but as the M-Code interpreter was designed to be run on
a 1980's ~5 MHz IBM PC it performs quite well enough on a today's 2500+ Mhz
systems ;-)

What I *would* be interested in is a copy of the 8088 Assembler source code
of that M-Code interpreter. The documentation indicates to me that the
programmers were Rodney Riggs and / or Rod Schiffman who I believe were both
either at, or recently graduated from, Brigham Young University in Utah at
the time.

--
Chris Burrows
CFB Software

[Fruttenboel]

Post by Chris Burrows
Why rewrite it in Oberon - there's plenty of Intel Modula-2 compilers
around. Then again, why bother to do that. The M2M-PC system runs the
original Lilith compiler and Medos on top of an M-Code interpreter which was
http://www.cfbsoftware.com/modula2

I found that yesterday after having posted my previous mail.. :o)

I would care to do some rewriting to circumvent DOS. You don't want MS
to come chase you for royalties afterwards. And it wouldn't be a real
Lilith on top of DOS.

For FST I wrote a simple graphic library a decade ago that can be used
as a starting point

http://fruttenboel.verhoeven272.nl/modula-2/vgalib3.html

Post by Chris Burrows
If the Lilith OS and applications were ported to an Intel Modula-2 compiler
(NOT a trivial task due to the differences of PIM 2/3/4 and ISO M2) it would
certainly run faster but as the M-Code interpreter was designed to be run on
a 1980's ~5 MHz IBM PC it performs quite well enough on a today's 2500+ Mhz
systems ;-)

Yeah, speed won't be a problem. It will even run nice in a Linux
dosbox.
For the differences between PIMs: just stick to PIM3. Forget ISO. ISO
is nice for nuts and bolts but useless for the rest.

Post by Chris Burrows
What I *would* be interested in is a copy of the 8088 Assembler source code
of that M-Code interpreter. The documentation indicates to me that the
programmers were Rodney Riggs and / or Rod Schiffman who I believe were both
either at, or recently graduated from, Brigham Young University in Utah at
the time.

I cannot imagine an M code interpreter to be that hard to write (for a
programmer in 2010) so i guess that would be the easy part...

I think it's time to dig up an old latitude to install DOS on it abd
see how far I will come...

[Chris Burrows]

Post by Fruttenboel
I cannot imagine an M code interpreter to be that hard to write (for a
programmer in 2010) so i guess that would be the easy part...

There's a 95% version written in Modula-2 in the appendix of "ETH Report Nr
40 - The Personal Computer Lilith". I used that to develop an MC68000
version that ran on an NCR Tower under Unix and also a JPI Modula-2 version
for MS-DOS. The more difficult part that didn't get completed at that time
was the linker / loader - would be much easier these days now that the
source code of Lilith and Medos is publicly available.

--
Chris Burrows
CFB Software
http://www.cfbsoftware.com/modula2

[Christoph Schlegel]

Post by Chris Burrows

Post by Fruttenboel
and, this opens an alternative: if one has Oberon0 booting, it should
be possible to rewrite Lilith in Oberon0 and some 386 assembler. just
put the dreaded x86 in flat real mode and you have the finest of
processors one can imagine.

Why rewrite it in Oberon - there's plenty of Intel Modula-2 compilers
around. Then again, why bother to do that. The M2M-PC system runs the
original Lilith compiler and Medos on top of an M-Code interpreter which was
http://www.cfbsoftware.com/modula2
If the Lilith OS and applications were ported to an Intel Modula-2 compiler
(NOT a trivial task due to the differences of PIM 2/3/4 and ISO M2) it would
certainly run faster but as the M-Code interpreter was designed to be run on
a 1980's ~5 MHz IBM PC it performs quite well enough on a today's 2500+ Mhz
systems ;-)

there is m2f which is a PIM2 compiler for Linux and similar systems
built using gcc.

Post by Chris Burrows
What I *would* be interested in is a copy of the 8088 Assembler source code
of that M-Code interpreter. The documentation indicates to me that the
programmers were Rodney Riggs and / or Rod Schiffman who I believe were both
either at, or recently graduated from, Brigham Young University in Utah at
the time.
--
Chris Burrows
CFB Software

[Chris Burrows]

Post by Fruttenboel
but, i got bored of oberon, being OS and compiler in one.

That appears to have been a common misconception - it's an option, not a
necessity. I assume you are aware that there have been a number of
implementations of Oberon compilers independent of the Oberon OS available
for at least 15 years? I have used three on Windows alone i.e. POW!, XDS and
Oberon/L (now Blackbox Component Builder).

--
Chris Burrows
CFB Software
Armaide: Oberon-07 Development System for ARM
http://www.armaide.com

[Fruttenboel]

Post by Chris Burrows
implementations of Oberon compilers independent of the Oberon OS available
for at least 15 years? I have used three on Windows alone i.e. POW!, XDS and
Oberon/L (now Blackbox Component Builder).

Windows... heared of that before... american, isn't it? from a small
time company somewhere on the west coast. never tried it though. don't
like these newish thingies that come and go.

hmm. perhaps i should take a look in that windows fling, if it doesn't
consume too many system resource.

[Fruttenboel]

Post by Chris Burrows
I have used three on Windows alone

as may have been clear from the other mail: I run Linux. And the 2.6
kernel developers forgot to keep the backward compatibility so now LNO
doesn't run anymore.

LNO was fine. it had all you could wish for.

http://fruttenboel.verhoeven272.nl/Oberon/index.html

[Chris Burrows]

Post by Fruttenboel
as may have been clear from the other mail: I run Linux. And the 2.6
kernel developers forgot to keep the backward compatibility so now LNO
doesn't run anymore.

I haven't tried it on Linux myself but Native XDS-x86 Modula-2 / Oberon is
available for Linux kernel 2.2 or higher:

http://www.excelsior-usa.com/xdsx86linux.html

--
Chris Burrows
CFB Software
http://www.cfbsoftware.com/modula2

[Olafur Gunnlaugsson]

Post by Fruttenboel

Post by Chris Burrows

Post by Fruttenboel
it would be better to rewrite the software for the pentium I processor
and the pc platform. better and cheaper.

Nah! What would the challenge be in doing that?

that's what the real men did with aos: port oberon to the dreaded x86
environment. all other platforms are irrelevant nowadays. x86 is the
borg. resistance is futile. they will all be assimilated.

there are more PowerPC systems sold in a day than Modula2/Oberon
development systems are sold in a year, even pascal is generating more
light than heat these days

the irrelevant thing is Wirth family languages I am afraid

Post by Fruttenboel
it will still be a hell of a job to make a stand alone lilith86
system. enough of a challenge for a mortal soul. at least, in the
northern hemisphere it would. :o)

[Chris Burrows]

Post by Jos Dreesen
Here are the Lilith schematics in Eagle format....
ftp://jdreesen.dyndns.org/ftp/CAD
If anyone is crazy and/or foolhardy enough to create new PCB's, then I want a set !

I've been thinking...

A year or so ago I built from scratch a simple 4K-RAM S2650-based
microcomputer that I originally had built for a kit some time around 1980. A
photo is here:

Loading Image...

At the time one of the options I considered was a hybrid solution where I
would use a genuine CPU but it would be interfaced to an ARM microcontroller
instead of the original discrete components needed to satisfy the ROM / RAM
requirements. The intention was to implement software for the ARM so that
the CPU data and address lines would receive the same electrical signals
from the ARM microcontroller as when connected to the discrete EPROM / RAM
chips and the 74LS138 memory address decoder.

The advantages of this approach would have been:

1. No need to source the old RAM chips
2. 32K of RAM (the CPU limit) could be accessed instead of the 4K limit of
the original design
3. No need for the piggyback adapter board which was required to substitute
the pin-incompatible EPROM for the unavailable monitor ROM.

Would a similar scheme be feasible to re-implement a Lilith? i.e. use a
genuine AMD 2901 processor with the micro-coded instruction set as the CPU
but implement as much of the rest of the system i.e. RAM, video, mass
storage, keyboard etc. using current parts (e.g. an LCD display with a smart
controller, SD card for mass storage using an SPI interface etc.)

Regards,
Chris

--
Armaide: ARM Oberon-07 Development System
http://www.armaide.com

[Fruttenboel]

Post by Chris Burrows

Post by Jos Dreesen
If anyone is crazy and/or foolhardy enough to create new PCB's, then I want a set !

I've been thinking...
A year or so ago I built from scratch a simple 4K-RAM S2650-based
microcomputer that I originally had built for a kit some time around 1980.
At the time one of the options I considered was a hybrid solution where I
would use a genuine CPU but it would be interfaced to an ARM microcontroller
instead of the original discrete components needed to satisfy the ROM / RAM
requirements.

Clever.

Post by Chris Burrows
Would a similar scheme be feasible to re-implement a Lilith? i.e. use a
genuine AMD 2901 processor with the micro-coded instruction set as the CPU
but implement as much of the rest of the system i.e. RAM, video, mass
storage, keyboard etc. using current parts (e.g. an LCD display with a smart
controller, SD card for mass storage using an SPI interface etc.)

yes it would/will.

but we should have learned from professor wirth: never do a double
development debugging drama!

you would need to tackle the software bugs in your ported code, next
to debugging the hardware design errors simultaneously. a stray beep
is simple. that's hardware. but a stray missed memory access can be
either hardware or software. you need lots of gear and time to test
that.

that's why i opted for the (truely tested) pentium board, like this
one:

http://www.asrock.com/mb/overview.asp?Model=A330GC

a stable platform for 70 euro's so you can concentrate on the
software. all you will ever need is on this board:

vga
dual sata drive
gbit lan
sound
the lot

no arm will beat this.

another option: http://www.asrock.com/mb/overview.asp?Model=N68-S 36
euro's plus a cpu (3 GHz Sempron) for 33 euro's and we're back at the
magic 70 euro's.

still, you won't be able to design and produce an arm board that is as
potent as this (yikes) pc board....

face it chris. you can run anything on an x86 board for almost free.
that's how the world is nowadays.

[tbreeden]

Post by Chris Burrows
I've been thinking...
At the time one of the options I considered was a hybrid solution where I
would use a genuine CPU but it would be interfaced to an ARM microcontroller
instead of the original discrete components needed to satisfy the ROM / RAM
requirements. The intention was to implement software for the ARM so that
the CPU data and address lines would receive the same electrical signals
from the ARM microcontroller as when connected to the discrete EPROM / RAM
chips and the 74LS138 memory address decoder.

This sounds a bit similar to the most recent Amiga "rebirth"
announcement, a
relatively high end PPC board will be integrated with one or more
XMOS
http://www.xmos.com/ programmable chips.

Details of the "Xorro" connection bus have not yet been made clear,
but it may
be possible to do something like you describe (otherwise, what's the
point?).
The XMOS core chips are really cheap, so it may be possible to do
interesting
things with a pile of them working together. (unfortunately, I'm not
familiar
enough at this level to know).

Anyway, it's just the thing to use with my Amiga PPC Modula-2 compiler
for
development - Aglet M2 - http://home.ntelos.net/~tbreeden/ :)

What's been revealed about this upcoming machine so far is at

http://www.a-eon.com/6.html

regards,
Tom

[Chris Burrows]

Post by tbreeden
This sounds a bit similar to the most recent Amiga "rebirth"
announcement, a relatively high end PPC board will be
integrated with one or more XMOS
http://www.xmos.com/ programmable chips.

Interesting ...

Coincidentally a discussion titled "using an FPGA to emulate a vintage
computer" has since started in the alt.folklore.computers newsgroup. One of
the references given there is "FPGA retrocomputing":

http://www.merlintec.com:8080/hardware/31

Post by tbreeden
Anyway, it's just the thing to use with my Amiga PPC Modula-2 compiler
for development - Aglet M2 - http://home.ntelos.net/~tbreeden/ :)

Always good to see new M2 compiler development. Your earlier Amiga 68000
Amiga compiler would run on a MiniMig wouldn't it?

http://en.wikipedia.org/wiki/Minimig

--
Chris Burrows
CFB Software
http://www.cfbsoftware.com/modula2

[tbreeden]

Post by Chris Burrows
Always good to see new M2 compiler development. Your earlier Amiga 68000
Amiga compiler would run on a MiniMig wouldn't it?
http://en.wikipedia.org/wiki/Minimig

It probably would, though I haven't looked into the Minimig.
It remains to be seen how useful it will be to have the XMOS chips as
part of the package.
People are saying that they are much easier to program than the FPGAs.
At $6 a core they sound good to me, but does that price them out of
the FPGA market?

Tom

[Olafur Gunnlaugsson]

Post by tbreeden

Post by Chris Burrows
Always good to see new M2 compiler development. Your earlier Amiga 68000
Amiga compiler would run on a MiniMig wouldn't it?
http://en.wikipedia.org/wiki/Minimig

It probably would, though I haven't looked into the Minimig.
It remains to be seen how useful it will be to have the XMOS chips as
part of the package.
People are saying that they are much easier to program than the FPGAs.
At $6 a core they sound good to me, but does that price them out of
the FPGA market?
Tom

XMOS is for totally different applications than a FPGA

FPGA's are expensive in comparison, but more flexible