Bit of CP/M trivia needed - test-drb@ccmp.vtda.org

30 Aug 2005

...

Subject: Re: Bit of CP/M trivia needed
   From: "Dwight K. Elvey" <dwight.elvey at amd.com>
   Date: Tue, 30 Aug 2005 18:04:53 -0700 (PDT)
     To: cctalk at classiccmp.org

 From: "Allison" <ajp166 at
bellatlantic.net> ---snip---

All the S100 machine of early origin were front pannel
(switches and lights) and few had rom/eprom. 
Hi Allison
 Yes, eproms were expensive. 
Ah just to set a point.  I still have the first 32byte 
fuseprom and 1702 I'd ever bought.  It's very painful 
for me to know how much I paid for them.  So for a fact
I do know they were expensive and also tough to get because
as both an engineer and hobbiest I was there buying them.

...
    Most all 
used hardware IO and a hand toggled in boot (assuming 
the controller didn't have a bootrom).  Most of the 
people doing it were running far lighter hardware 
than you have there.  Most of the heavey geeks were 
using stuff like PT ALS8 systems to bootstrap to 
CP/M for the first time.  Me I plodded through a 
lot of hand toggling plus my tape based system. 
 I did this also to get my first level of bootstrap
running and saved to the disk. I'll admit that I did
take advantage of having a PC laptop to use as
large external data storage to minimize the amount
of toggling. Still, I made a formatting program,
first level bootstrap and BIOS through the front panel
with the serial input. I used no monitor program,
just a simple serial-input-to-memory program that I
toggled in.
 I can imagine how much more effort there would have
been without the external mass storage. Even a minimum
of toggling and paper tape would be quite painful. 
That not even the first level guess.  It was harder because 
there was little experience out there to draw from and if 
you had a disk early on likely you were one of the 
few.  So there was a lot of by the seat of the pants
and sweat to make things work. I'm talking prior to 1977
as that summer a lot more system with matching disks 
started to appear (NS* was one) and the market was ready.

...
  Even then it
was load the available image on SSSD 8"
then overlay the drivers by hand (switches or monitor 
program) then save it to whatever before a trial boot.
Early on there were very few disk systems and fewer 
that were plug and go for CP/M.  Usual case was 
migrating from available hardware and non-CP/M software
to CP/M.  Did that for both 8" and NS* 5.25 and also the 
NEC PDA-80 I had.  To this day I cringe when I do a 
first boot even thugh I have moden tools and usually 
boot from EPROM instead.  Such is memory of the process. 
 You always wonder if the program you just toggled in
is going the wipe out all the work you previously
entered. Or worse, wipe out a CP/M distribution disk. 
!$%^&, Now I"m going to have them nightmares again.
Really, it happened too often.

The day I put in the NS* controller toggled E900h 
(ok it was 351,000 split octal) into the Altair front 
pannel hit loadaddr then run was like I could hear a 
chorus. Boots thereafter were almost reliable. Once the 
NS* chassis arrived the Altair was retired and reliability
was established.

...

It wasn't until around late '77 (same time I got my NS*)
that I started seeing turnkey machines as the norm.   
 As you state, normal turnkey used boot ROMs/EPROMs. One
might consider my setup to be turnkey but it isn't a
normal turnkey. There is no boot ROM/EPROM. It just uses
the normal reset of the controller without main CPU
interaction. It is not fully automatic since I do need
to hit the reset and wait for the disk to complete before
enabling the computer to go. Almost full turnkey.
Dwight 
==An aside from the questio to set background=============

Two address history you have to understand the enviroment
as many experienced it.  Cheap computers then was any
machine under 5000$.  Back then $5000 was a good down 
payment on a house in the northeast.  

The significance is that your hardware was not typical
nor commonly seen.  Most of the people didn't have that 
kind of money and likely before early '77 it didn't exist.
More typical was a PT SOL/W NS* or NS*, SWTP 6800 systems
and a few maybe of the Altair or Imsai style plus a few 
bare bones scratch builts.   I attended more than a few 
LICA (Long Island [NY] Computer Assoc) meetings where 
the common machine was an ELF more maybe an AMD EVK68 
(6800 SBC), Moto 6800D1 or National SC/MP.  I was 
somewhat of an oddity as I had by then Altair, SC/MP, 
ELF and even a Technico Super Starter board (TI9900)
by time the TRS80 was just appearing.  In the whole 
association I think there were maybe 8 of maybe 100+
that even had a working disk.

Just so you have an idea.  If you had a disk in early 
1977 and ran CP/M likely your outlay in Dollar$ was 
in the range of 4500-6000+(US$) into it and that didn't
include a terminal or TTY (add ~1000$).  Around then
a new Chevy C20 4wd truck was 3800$ for comparison.

To run CP/M 1.3 or 1.4 you needed a minimum of 16k 
of ram, a disk with controller, IO board, terminal
and an 8080.

In 1976 summer: Some of the 1977 list items were
unotainium. Those items that could be bought were
roughly 2x the 1977 price.  IE 16K of ram was 
$1000.

1977 summer cost:
Barebones Altair or Imsai ($395 kit)
SA400 (not the 400L) $399 new 
8K ram board seals/PT/IMS $259(kit)
Floppy controller ($699[kit] with 1st drive NS*)
ICOM FDC and disk(sa400) $1095
Z80 CPU (NS* $199)
Serial IO card 159$
Tarbel cassette interface $120(kit)
TTY used $800-900, VDM-1 199(kit)
Printer (see tty)

In 1978 that stuff would be around half the price.

In late 1978 I took a Data structures in pascal course 
in college at night.  Of the professors 200+ students
I was one of 6 that owned a computer (not a pocket calc)
of some sort and one of two that had a disk and could
run UCSD pascal. Back then a personal computer was 
_any_ hardware that was owned personally.  

========================================================

Allison