18 Nov
2003
18 Nov
'03
7:29 p.m.
On 11/18/2003 01:23 AM -0600, Tom Jennings <tomj(a)wps.com> wrote:
Tom wrote lots of good things...
Some BOZO designer had, I assume arrogantly, put all
the video
interrupts (vert, horiz retrace, etc) starting at 00020h, which everyone
knows is where MSDOS has it's software interrupts. They were in a PAL or
something, and could not be moved.
Well, I won't defend the quirks of the Rainbow... (more later)
But you're forgetting that the Rainbow 100A design pre-dates the IBM PC, it
was first designed as a CP/M machine. MS-DOS came later.
And IBM put interrupts on the PC in places that Intel's data sheets
documented as RESERVED! IBM did not pay attention to Intel's
recommendations and violated a few of their support device allocations.
DECnet-DOS was developed in the DECnet group inspite of the Rainbow group
at first. Their marketeer felt that Corvus Omninet was going to be the big
thing. Only later did we integrate when they got into LAN Manager support.
Our first target was the IBM XT with a 3Com 3C501 Ethernet board, and the
Rainbow running DDCMP on the serial port. We weren't going to wait for
the Rainbow-150 which would be a more industry compatible system. When the
IBM AT appeared, that scotched the RB-150 and lead, eventually... to the
VAXmate. (Thank KO for the delays)
I remember the MHFU watchdog and the Vertical Retrace Interupt. The DECnet
kernel basically ran as an interrupt driven background TSR. On the IBM PC
we easily shelled off the 1C clock tick, but on the Rainbow, we had to
build "clever" reentrancy shells around several interrupts to avoid some
wierd stack overflow or reentrancy into the video BIOS.
The Rainbow at least had a decent UART with a multicharacter FIFO. I had
no problem doing 19.2kbs using C routines. That stupid XT UART could
barely do 9600 with an optimized state driven assembly routine. UART FIFOs
didn't appear until the IBM PS/2 Model 50.
Dave (has two HP16Cs and runs the wprn program)
19 Nov
19 Nov
1:09 p.m.
On Tue, 2003-11-18 at 22:29, Dave Mitton wrote:
My apologies to the engineer, I should not have insulted him/her even
Some BOZO
designer had, I assume arrogantly, put all the video
interrupts (vert, horiz retrace, etc) starting at 00020h, which everyone
knows is where MSDOS has it's software interrupts. They were in a PAL or
something, and could not be moved.
Well, I won't defend the quirks of the Rainbow... (more later)
But you're forgetting that the Rainbow 100A design pre-dates the IBM PC, it
was first designed as a CP/M machine. MS-DOS came later. from afar, I have no idea what their criteria might
have been, but I
still maintain it was very short sighted. I assume it was decreed
from
'above'.
Whether MSDOS 'winning' (sic) the OS war of the time (vs. CP/M-86) was
obvious or not, MSDOS was in no way ignorable. The 8088 card added to
the Rainbow design post-dates MSDOS (and certainly post dates 86DOS,
which MSDOS 1.25 is, bought from Seattle Computer Products (man I wish I
had saved those 86DOS 0.86 and 1.x manuals!), which *someone* at DEC
must have known about. And if they didn't they weren't paying attention.
It's hard to imagine they chose interrupts 0020h - 002fh by accident;
it's just Good Practice to avoid all possible conflicts, especially when
the cost is essentially zero to do so.
And IBM put interrupts on the PC in places that
Intel's data sheets
documented as RESERVED! IBM did not pay attention to Intel's
recommendations and violated a few of their support device allocations.
Yeah, they were stupid too. I've got two words for the IBM PC ROM code:
ug lee. The biggest steamroller wins.
I remember the MHFU watchdog and the Vertical Retrace
Interupt. The DECnet
kernel basically ran as an interrupt driven background TSR. On the IBM PC
we easily shelled off the 1C clock tick, but on the Rainbow, we had to
build "clever" reentrancy shells around several interrupts to avoid some
wierd stack overflow or reentrancy into the video BIOS.
Ahh... now I get it. I wasn't told even this much back then!!
I do recall some severe stack-depth problem, I had to switch stack
pointers in the I/O drivers somewhere I recall. I think. Wish I had the
code!
The Rainbow at least had a decent UART with a
multicharacter FIFO. I had
no problem doing 19.2kbs using C routines. That stupid XT UART could
barely do 9600 with an optimized state driven assembly routine. UART FIFOs
didn't appear until the IBM PS/2 Model 50.
Yes, when DEC applied themselves without weird non-functional agendas
they made really nice stuff. It made a nice Fido bulletin board (umm but
not the 100A's... :-(
1:19 p.m.
On 19 Nov 2003, Tom Jennings wrote:
I do recall some severe stack-depth problem, I had to
switch stack
pointers in the I/O drivers somewhere I recall. I think. Wish I had the
code!
[mumbles something about silly programmers and not making
backups]
--back to converting PS docs to PDF, and checking 'em
--
Fred N. van Kempen, DEC (Digital Equipment Corporation) Collector/Archivist
Visit the VAXlab Project at http://www.pdp11.nl/VAXlab/
Visit the Archives at http://www.pdp11.nl/
Email: waltje(a)pdp11.nl BUSSUM, THE NETHERLANDS / Sunnyvale, CA, USA
2:17 p.m.
On Wed, 2003-11-19 at 16:19, Fred N. van Kempen wrote:
On 19 Nov 2003, Tom Jennings wrote:
It gets worse!
I *did* backups. Now I certainly was dumb enough to delete old BIOS code
(there were also liability issues; it wasn't my code), but here's the
worst part:
I had tape backup starting in the CP/M days (I wrote a start/stop
DC300-type tape backup system for Alloy), went without for a small gap
in the pre-hard-disk days of MSDOS (when 10MB RLL MFM drives cost actual
money), then QIC and other formats, etc.
So I'm merrily doing monthly dumps, rotating weekly dumps and often
daily diff tapes, and got lax in checking the tapes afterwards... I
didn't realize my tape drive had been drifting out of alignment(or
something) until the 100MB hard disk crashed suddenly, and found I had
lost my life's work. This was in 1993 or 94.
I no longer do "backups". I take John Gilmore's advice, everything is on
a rotating spindle in a machine on the net running syslog, etc. If one
of the drives (etc) goes bad, it gets immediate attention like any
operating computer does. Cron and rsync and ssh, machines separated 1500
miles apart. Four live, error-corrected copies of everything. Plus a few
CDROM copies (as universal a medium as IBM3740, I'm sure).
Plus I'm slowly printing on paper. The only stuff I have left that's
"old" is 6800 code I wrote for my SWTPC, a 7-track magtape filesystem
and executive (it's godawful, filled with embarrassments and horrors,
but it was interrupt driven, very clever) which was printed on good old
lineprinter paper.
I can take a hint.
Use good paper, good ink.
Maybe I shoudl punch a few thousand kilometers of tape?
I do recall some severe stack-depth problem, I
had to switch stack
pointers in the I/O drivers somewhere I recall. I think. Wish I had the
code!
[mumbles something about silly programmers and not making
backups] 
7:39 p.m.
So with all this talk to Rainbows is there a really good site with lots of
photo's, memo's and information on the whole creation, design and technicals
of the Rainbows??? If not, there should be! ;-)
Curt
----- Original Message -----
From: "Gene Buckle" <geneb(a)deltasoft.com>
To: "General Discussion: On-Topic and Off-Topic Posts"
<cctalk(a)classiccmp.org>
Cc: "General Discussion: On-Topic Posts Only" <cctech(a)classiccmp.org>
Sent: Wednesday, November 19, 2003 8:53 PM
Subject: Re: Dec Rainbow prehistory
Maybe I shoudl punch a few thousand kilometers of tape?
....make sure it's mylar....and stored in a controlled temperature, argon
filled box. :)
BTW, I'd love to get my hands on the FidoNet code if you've got any left.
*grins*
g.
20 Nov
20 Nov
1:14 a.m.
-----Original Message-----
From: cctalk-bounces(a)classiccmp.org
[mailto:cctalk-bounces@classiccmp.org]On Behalf Of Curt vendel
Sent: 20 November 2003 06:40
To: General Discussion: On-Topic and Off-Topic Posts
Subject: Re: Dec Rainbow prehistory
So with all this talk to Rainbows is there a really good site with lots of
photo's, memo's and information on the whole creation, design and
technicals
of the Rainbows??? If not, there should be! ;-)
I'd love to bolt it onto my needing-fleshing-out Digital pages at Binary
Dinosaurs.....
cheers
--
adrian/witchy
www.binarydinosaurs.co.uk - the online computer museum
www.snakebiteandblack.co.uk - ex-monthly gothic shenanigans :o(
8:46 a.m.
On Wed, 2003-11-19 at 17:53, Gene Buckle wrote:
ANother atrocity, blame entirely on me: version 11w sources, the 'best'
pre-zone version, lost. Final release 12something, also lost due to my
carelessness.
However, a post-12 set of souces is on my website, wps.com, look in the
table of contents. It compiles, almost links (Microsoft C). I prototyped
it all (was originally K&R C, Lattice compiler!). It used PLINK and
their small-model overlay loader to run in 640K. Good luck finding that.
Maybe I shoudl punch a few thousand kilometers of tape?
....make sure it's mylar....and stored in a controlled temperature, argon
filled box. :)
BTW, I'd love to get my hands on the FidoNet code if you've got any left.
*grins* 
9:50 a.m.
Maybe I shoudl punch a few thousand kilometers of
tape?
....make sure it's mylar....and stored in a controlled temperature, argon
filled box. :)
BTW, I'd love to get my hands on the FidoNet code if you've got any left.
*grins* However, a post-12 set of souces is on my website,
wps.com, look in the
table of contents. It compiles, almost links (Microsoft C). I prototyped
it all (was originally K&R C, Lattice compiler!). It used PLINK and
their small-model overlay loader to run in 640K. Good luck finding that.
I've actually got most or all of those tools. I got a CD with the entire
OpenDOS (recently renamed from Novell DOS 7 in 1996) build environment
including all the tools. It included at least 2 different versions of
Lattice C. I think I saw PLINK in there too, but it's been at least 5
years since I've look at that CD.
Regardless, it's damn gcool that you've got it online. Thanks.
Now, how do I get Wynn Wagner III to cut loose the Opus source? *grin*
g.
22 Nov
22 Nov
7:25 a.m.
On 20 Nov 2003, Tom Jennings wrote:
However, a post-12 set of souces is on my website,
wps.com, look in the
table of contents. It compiles, almost links (Microsoft C). I prototyped
it all (was originally K&R C, Lattice compiler!). It used PLINK and
their small-model overlay loader to run in 640K. Good luck finding that.
I have PLINK on some old disk laying around somewhere, though it will take
a LOT to compel me to go looking for it in case anyone is interested.
--
Sellam Ismail Vintage Computer Festival
------------------------------------------------------------------------------
International Man of Intrigue and Danger http://www.vintage.org
[ Old computing resources for business || Buy/Sell/Trade Vintage Computers ]
[ and academia at www.VintageTech.com || at http://marketplace.vintage.org ]
19 Nov
19 Nov
1:24 p.m.
It was thus said that the Great Tom Jennings once stated:
On Tue, 2003-11-18 at 22:29, Dave Mitton wrote:
My apologies to the engineer, I should not have insulted him/her even
from afar, I have no idea what their criteria might have been, but I
still maintain it was very short sighted. I assume it was decreed from
'above'.
Why was it short sighted? Intel reserved all interrupts between 0 and 31,
leaving 32 (or 20h) to be the first interrupt available for programmer or
system use. Since the design was before MS-DOS, it doesn't seem all that
far fetched to start the hardware interrupt vectors at 20h. If anything,
the engineer was following the guidelines set by Intel; IBM, on the other
hand, ignored Intel.
Some BOZO
designer had, I assume arrogantly, put all the video
interrupts (vert, horiz retrace, etc) starting at 00020h, which everyone
knows is where MSDOS has it's software interrupts. They were in a PAL or
something, and could not be moved.
Well, I won't defend the quirks of the Rainbow... (more later)
But you're forgetting that the Rainbow 100A design pre-dates the IBM PC, it
was first designed as a CP/M machine. MS-DOS came later. Whether MSDOS 'winning' (sic) the OS war of
the time (vs. CP/M-86) was
obvious or not, MSDOS was in no way ignorable. The 8088 card added to
the Rainbow design post-dates MSDOS (and certainly post dates 86DOS,
which MSDOS 1.25 is, bought from Seattle Computer Products (man I wish I
had saved those 86DOS 0.86 and 1.x manuals!), which *someone* at DEC
must have known about. And if they didn't they weren't paying attention.
It's hard to imagine they chose interrupts 0020h - 002fh by accident;
it's just Good Practice to avoid all possible conflicts, especially when
the cost is essentially zero to do so.
When was the Rainbow introduced? MS-DOS (PC-DOS) 1.0 wasn't introduced
until August of 1981, and it was in August or September of 1980 that
Microsoft licensed QDOS from Seattle Computer Products; I'm not sure when
QDOS was first introduced, but my guess is around 1979 or there-abouts.
Also, DEC was primarily a minicomputer manufacturer so it may not be that
far fetched that engineers there might not be terribly familiar with state
of the art (hah!) microcomputer products at the time.
-spc (When will this stuff dislodge from my brain? 8-)
2:29 p.m.
On Wed, 2003-11-19 at 16:24, Sean 'Captain Napalm' Conner wrote:
re: int 20h
Why was it short sighted? ... Since the design was
before MS-DOS,
Oh oh... was it really? Is it possible that my mental chronology is
corrupted? Gee, that's never happened before, erm, ugh...
when was the rainbow produced?
I don't recall exactly, I'd guess 1981 or 1982.
I'm not sure when
QDOS was first introduced, but my guess is around 1979 or there-abouts.
Obviously I may be utterly wrong here then.
(I recall the QDOS name also, barely, but it was also called 86DOS at
some point; what was what where?)
Also, DEC was primarily a minicomputer manufacturer so
it may not be that
far fetched that engineers there might not be terribly familiar with state
of the art (hah!) microcomputer products at the time.
Well... I lack sympathy here, still; obviously they were building
machinery for this realm, it was beholden on them to know, as much as
was possible then of course. DEC was very arrogant about CP/M-86 being
superior; they told me as much many times.
-spc (When will this stuff dislodge from my brain?
8-)
Never, it just changes shape :-)
2:31 p.m.
It was thus said that the Great Tom Jennings once stated:
I'm not sure how system calls are made in CP/M-86, but I do know that
there were two ways to call DOS functions, via INT 21h, and through a call
to location $005 (in the PSP segment), which emulates the CP/M system
calling convention (which I think just did an INT 21h/RET, although
gratefully the memory is fading on that stuff).
Heh.
-spc (Mommy! Make it go away!)
On Wed, 2003-11-19 at 16:24, Sean 'Captain Napalm' Conner wrote:
re: int 20h
If it was introduced in 1981, I would say work was done at the same time
that the IBM PC and MS-DOS were being worked on; if around 1982 it might
have been the design itself predated MS-DOS but that any changes to the
system might be too expensive and the designers felt it would be an easy
thing to work around in software.
Why was it short sighted? ... Since the
design was before MS-DOS,
Oh oh... was it really? Is it possible that my mental chronology is
corrupted? Gee, that's never happened before, erm, ugh...
when was the rainbow produced?
I don't recall exactly, I'd guess 1981 or 1982. Also, DEC was
primarily a minicomputer manufacturer so it may not be that
far fetched that engineers there might not be terribly familiar with state
of the art (hah!) microcomputer products at the time.
Well... I lack sympathy here, still; obviously they were building
machinery for this realm, it was beholden on them to know, as much as
was possible then of course. DEC was very arrogant about CP/M-86 being
superior; they told me as much many times. -spc (When
will this stuff dislodge from my brain? 8-)
Never, it just changes shape :-)
20 Nov
20 Nov
8:43 a.m.
On Wed, 2003-11-19 at 17:31, Sean 'Captain Napalm' Conner wrote:
I'm not sure how system calls are made in
CP/M-86, but I do know that
there were two ways to call DOS functions, via INT 21h, and through a call
to location $005 (in the PSP segment), which emulates the CP/M system
calling convention (which I think just did an INT 21h/RET, although
gratefully the memory is fading on that stuff).
Ugh! I hate you! I had forgotten that, and you refreshed the memory! :-)
Only a small subset of the MSDOS calls were available via 0100:0005[], I
believe. Oh the horror, the horror, I'll probably wake up from a dream
trying to code a PUNCH call from Perl or something. Where's that damned
IOBYTE?
19 Nov
19 Nov
2:13 p.m.
At 04:00 PM 11/19/03, you wrote:
On Tue, 2003-11-18 at 22:29, Dave Mitton wrote:
Whether MSDOS 'winning' (sic) the OS war of the time (vs. CP/M-86) was
obvious or not, MSDOS was in no way ignorable. The 8088 card added to
the Rainbow design post-dates MSDOS
um, I have rainbow A & B models, and the 8086 is on the main logic board,
not on an add on card.
20 Nov
20 Nov
8:40 a.m.
On Wed, 2003-11-19 at 17:13, Geoff Reed wrote:
um, I have rainbow A & B models, and the 8086 is
on the main logic board,
not on an add on card.
Sorry for my imprecision. You are right, it wasn't an "add-on" card, the
context for the incorrect remark was the Rainbow described as a CP/M
machine, with the 8086 "added on" to it. Obviously not so literally as I
said.
(Logically-speaking, the 8086 was an accessory to the Z80, whivch owned
all the peripherals.)
9:37 a.m.
At 11:30 AM 11/20/03, you wrote:
On Wed, 2003-11-19 at 17:13, Geoff Reed wrote:
I always found the Rainbow architecture kinda cool actually, when running
in x86 mode the Z-80 functioned as a dedicated IO processor, freeing the
X86 to run your programs without having to deal directly with the preipherals.
um, I have rainbow A & B models, and the 8086
is on the main logic board,
not on an add on card.
Sorry for my imprecision. You are right, it wasn't an "add-on" card, the
context for the incorrect remark was the Rainbow described as a CP/M
machine, with the 8086 "added on" to it. Obviously not so literally as I
said.
(Logically-speaking, the 8086 was an accessory to the Z80, whivch owned
all the peripherals.)
11:39 a.m.
On Thu, 2003-11-20 at 12:37, Geoff Reed wrote:
>(Logically-speaking, the 8086 was an accessory to
the Z80, whivch owned
>all the peripherals.)
I always found the Rainbow architecture kinda cool
actually, when running
in x86 mode the Z-80 functioned as a dedicated IO processor, freeing the
X86 to run your programs without having to deal directly with the preipherals.
If I recall, performance wise they werent' bad! They had shared
*memory*, not all that much of it and weird semaphores I think, but it
suuure beats a shared IO port or other nasty hack. Actual overlapping
IO.
It wasn't possible with MSDOS, but you could easily queue up disk
writes/reads on hte Z80 side, and have the Z80 tell the 8086 only upon
error, etc.
12:07 p.m.
Tom Jennings wrote:
If I recall, performance wise they werent' bad!
They had shared
*memory*, not all that much of it and weird semaphores I think, but it
suuure beats a shared IO port or other nasty hack. Actual overlapping
IO.
Was not a version of DECMATE? ( PDP-8 ) also a CMOS PDP-8/Z80 dual machine?
It wasn't possible with MSDOS, but you could
easily queue up disk
writes/reads on hte Z80 side, and have the Z80 tell the 8086 only upon
error, etc.
I still think the 8086 needed to have the 8080 instruction set microcoded
in and better 16 bit opcode design.
Ben.
5:13 p.m.
On Fri, 2003-11-21 at 10:07, ben franchuk wrote:
Was not a version of DECMATE? ( PDP-8 ) also a CMOS
PDP-8/Z80 dual machine?
There were (are?) add on processor boards for DECmate-IIs. There's
certainly a Z-80 one and I think there was also an 8086 one as well
(although I could be delusional here :-)
--
Huw Davies | e-mail: Huw.Davies(a)kerberos.davies.net.au
Melbourne | "If soccer was meant to be played in the
Australia | air, the sky would be painted green"
21 Nov
21 Nov
6:21 a.m.
On Thu, 20 Nov 2003, ben franchuk wrote:
Was not a version of DECMATE? ( PDP-8 ) also a CMOS PDP-8/Z80 dual machine?
Well, kind of. The DECmate II (and, I believe, the III; I never used a
III) could accept a coprocessor containing a Z80. The only I/O resources
directly available to the Z80 were registers used to pass messages
between itself and the PDP-8; the PDP-8 did all the I/O.
While I had to give up on the Rainbow because it was not good enough as
either an MS-DOS or a CP/M machine, the DECmate II is one of my all-time
favorite CP/M machines.
I knew the Rainbow was in trouble as an MS-DOS machine when I had to
trade my newer version of Turbo C for a friend's older version because
the *command line* version of the compiler had too many dependencies on
the IBM PC ROM BIOS. It was all down hill from there.
As a CP/M machine, I had to give up on it primarily because the Z-80
didn't support the I/O byte. At the time, the sort of dorking around I
was doing with CP/M involved serial communications of some sort; playing
about with serial mice, drawing things on a Tek 4010, communicating
with the VAX on which I did serious work, etc. All of these were done by
fiddling with the I/O byte on all of my CP/M machines *except* the
Rainbow. Consequently, I could write a program and use it on my
Kaypro, Epson Geneva, even the DECmate II; but *not* the Rainbow. I even
had a two-user experimental MP/M setup in which the XIOS sat beneath the
CP/M BIOS using the CP/M BIOS services for all I/O, which included
diddling the I/O byte to perform console I/O for the two users. Didn't
work on the Rainbow.
I never got attached to CP/M-86 because I didn't have development tools
or programming documentation for the system. In fact, my primary method
of using the Rainbow was to load my CP/M-80 TSR that gave me command
line recall and editing, which meant that from then on I could only run
Z80 programs. Consequently, the Rainbow's deficiencies as a CP/M-80
machine were fatal for me.
Oh yeah. The console I/O on the Rainbow was also slow. Not as bad as the
Pro, but the DECmate II's built-in terminal emulation was peppy.
--
Roger Ivie
rivie(a)ridgenet.net
<input type crash>
11:35 a.m.
A non-text attachment was scrubbed...
Name: not available
Type: text
Size: 1694 bytes
Desc: not available
Url :
http://www.classiccmp.org/pipermail/cctalk/attachments/20031121/e1e9916c/at…
7548
days inactive
7551
days old
22 comments
14 participants
participants (14)
-
ard＠p850ug1.demon.co.uk -
bfranchuk＠jetnet.ab.ca -
bob_lafleur＠technologist.com -
curt＠atarimuseum.com -
dave＠mitton.com -
geneb＠deltasoft.com -
geoffr＠zipcon.net -
huw.davies＠kerberos.davies.net.au -
rivie＠ridgenet.net -
spc＠conman.org -
tomj＠wps.com -
vcf＠siconic.com -
waltje＠pdp11.nl -
witchy＠binarydinosaurs.co.uk