Basic/Four Corporation Model 7250 terminal
jwstephens at msm.umr.edu
Fri Dec 15 23:47:47 CST 2006
Jay West wrote:
> Jim wrote....
> Oh wow. I did not know any of that, and I was around the early pick
> community and developers quite a bit. Fascinating. I didn't know the
> tie-in between MAI and Microdata. I knew the salesman we had for years
> selling pick stuff originally came from MAI, but I had no clue MAI had
> that common ancestry with Pick.
MAI had common ancestry with Microdata as did Pick. I meant that they
had a related approach in that once they built up the abstract model of
a basic language, file structure, etc, only a few programmers concerned
themselves with the "system" and the rest were applications people who
did MAI basic. Pick had their system people who concerned themselves
with the Pick virtual, OS, etc, then there is the legion of Pick
applications programmers who are still out doing pick database and
basic, even if a large number of them don't know it.
The interesting thing about the two models is that though pick was not
an "open "product as in todays open source model, there were around 25
to 30 entities who had ports of pick and had people who did the system,
and got more ideas than basic 4 who kept a lot of the applications and
system programming internal. I don't know if MAI supports any OEMs now,
and most of their current business is in the hospitatlity industry and
hotels, etc. Pick is probably #2 behind DB2 or maybe DB2 and
(wh)Oracle, and the largest implementation is owned by IBM (Unidata /
> They also ran on M1600's?
They were one of the first users.
> Holy cow... small world! So... did the company that made M1600's only
> target people who wanted to OEM their boxes with their own OS?
Microdata targeted not so much the OS as much as a large version of what
you see in Nuts 'n Volts today, the large OEM integrators who wanted to
have a general purpose computer to do something in an overall solution.
And the other contenders of the day, DEC, DG, General Automation, CA,
(insert list of OEM mini companys here) didn't fill the bill.
> Or did they have their OWN o/s? I'd never heard of anything on an
> M1600 besides Reality.
I wrote an OS we called MPS at UMR that ran on the 1600 you bought. I
used it at microdata for my own amusement for a couple of years, then
they started using it to support 1621 programming, since all their
hardware diagnostics were written in 1621 assembler, and I of course had
thoughtfully ported them all to run under MPS for both the assembly and
debugging. They had been using tty (TOS) and either cross assembly, or
later used a debugger that talked to port 0 of a reality system.
the parentage of the Microdata is interesting because you not only have
the OS but the firmware to deal with. The first firmware that was
written ran on the 800 hardware and was called the 810. The 800
firmware would run with minor tweaks on the 1600, so the 810 became the
1610. there were various extra instructions added, and there was
firmware called 811, 812, 813, and 821, all of which had 1600
counterparts. Someone at some later date rewrote the 821 and added some
stuff and came up with the 1630 which had enough stuff in it to not have
an 800 version. The 810 firmware was modeled similar to the Data
Machines 520, and have ex Data Machines engineers involved in the
design, hence the commonality.
Some time early on, Dick Pick and Chandru Murthi were shopping for a
platform to take the GIRLS database off the mainframe onto, and thru
Alan Shapiro, and Max Malone at microdata came up with using the 800.
If I am not mistaken, Republic Data (a tiny weird divison of Republic
Steel I think) contracted and they built up several 800's with 4k of
memory, 1 8 way, and a 5mb disk and a tape drive to run 8 tty's on an
800. Not bad if you think 1972 and the amount of hardwre involved. it
was reality 1.0 and had major problems, but was developed over several
years and morphs into the 1600 and Reality.
Basic 4 at some point, in the early 70's took a version of the 810 as
near as I can tell, and modified it to run a mixture of the 810
assembly, and a compiled basic with multi user enhancements. The 810
..thru 1630 were all classic "single user" systems, and had poor
architecures for multi users, but was good for running dedicated
systems. Basic 4 cleaned up things to allow for general purpose
timeshareing and took off from there.
At various times, Microdata sold systems to GTE that ran 1621, and
several other companies.
ADP bought a lot of reality systems and probably is the real money
engine that made pick, and at one time or another accounted for probaby
as much as 80% of the reality sales for microdata, and therefor also for
pick early on.
I know of oem systems that microdata sold that ran Harris firmware,
which was a cpu that Harris had developed. There was hardware to allow
the 1600 to run harris I/O as well as a control panel that matched the
harris cpu front panel, but was driven by firmware.
Another vendor bought and used 1600's running Varian Data Machines 620's.
I don't know if anyone ever used it commercially, but there was an 1130
emulation that ran, though I only ever saw the firmware. I never did
see any software that ran on it, so I don't know if it as more than an
exercise by the author, Cliff Myers. He wrote it while at Purdue, on a
system they had in their computer science department. I know of Purdue,
UMR, and University of Southwestern Lousianna (USL) that had Microdata
machines. The later had a 3200 or 32s and that machine is at the CHM,
according to Al Kossow.
So the OS question takes a bit to answer. Microdata officially only
ever supported an assembly kit that you could run from tty's for the
1600. That was the reason we wrote our Disk OS, MPS but it never was
officially acknowleged by Microdata as an "OS". the main system was
Reality, once it took hold and the OEM sales sort of withered. Also
Microdata had pursued a business model of "one stop buy everything" and
had invested heavily in loosing operations to manufacture OEM Tape
drives, Disk drives, terminals, and small printers. Only the use of the
tapes, disks and terminals made them even slightly a profit center.
Very few other companies I know of bought much of anything of the
Microdata peripheral line. And as the 1600 Realitys aged and microdata
didn't update the peripherals, they gradually got outflanked by
companies which could use cheaper and better tapes, disks and
peripherals. You Jay, will remember and shudder and heave if you recall
the Microdata competitor to the SMD world, the Reflex I and II. that is
why they eventually lost out to Ultimate IMHO.
> I did ask our salesman one day about MAI's systems, he said it was a
> business basic, but nothing at all like Pick (he wasn't technical
> though). Most strange. Thanks for the education :) '
I guess I misspoke as far as the commonality. I think that it did
resemble more the old Basic you saw on the 1600 than the Fortran like
Data BAsic that pick eventually supported. Pick's basic resembled a
sort of a Pascal with very bad design choices than a Basic language.
things like "let" was tolerated, but not required, and the syntax
gradually drifted towards procedural languages from the basic model.
Internally Basic 4 started out with a 1621 like firmware which ran the
OS in 32k (lower half of mem) and swapped multi users into the upper 32k
of a 64K 1600. It had a very bad 32k addressing design bug (the
designers used bit 15 of a 16 bit pointer to include the X (index) in
the address math in some cases, which poisoned it for any addressing,
hence only 32k jumps, unless you give up X. I think the nova did
something like that as well, but had a better way to do 64k or > 32k
object than the 1621 or 1630 did. Later versions got away from this
start, but preserved the business basic source.
> I can't wait to start restoration of my microdata M2000 :)
is that the microprocessor version, or is it one of the bigger reality
machines? don't recall right now. I left about when they came out with
the Sequel, and called it the M9000 (6000-8000 were reality 3.0 and
reality 4.0, so reality 5.0 ran on the 9000). after Microdata was
merged with the CMC mob, and they came out with things like the Spirit,
I lost track of all the model numbers, etc. But the M1000 was a box
oem'ed from Convergent technologies that microdata wedged a modifed
version of REvelation onto. Really didnt do much but modify the basic
compiler to accept some crap, and make it run on the Convergent Dos I
think. Convergent also ran a Unix, but microdata didn't buy that one.
I think they also ran on the X86 version of Convergent, not the 68000's
that they had later that scaled up into larger systems.
> Jay West
More information about the cctech