IBM 1130 Development

[garypw...@hotmail.com]

I was part of the original IBM 1130 operating system team.  It is interesting to hear about the things that are still being done with regard to the 1130.

[Bob Flanders]

Very cool.

What aspects did you work on?

Any stories to share?

Best regards (and a tip of the hat),
Bob Flanders

[Richard Stofer]

Your team did fantastic work. There are so many features it is hard
to single out any one particular item but LOCAL and SOCAL have to be
near the top of the list - very clever. I guess I still find it
staggering when I consider the level of complexity and the time in
history.

Your software lives on; I run it whenever I play with my FPGA 1130.
Just the way your team meant for it to be run.

Richard


On Jun 20, 7:35 am, "garypwhee...@hotmail.com"

[Paul Anagnostopoulos]

Oh, you have to tell us about the parts you worked on.

~~ Paul

[garypw...@hotmail.com]

I wrote the original Assembler, some of the I/O subroutines, and some
parts of DUP. Later, I managed the team when the macro capability was
added to the Assembler, and when RPG was integrated into DM2

On Jun 20, 8:22 am, Paul Anagnostopoulos <pcanagnostopou...@gmail.com>
wrote:

[Bob Flanders]

Aside from a user's request (copy a 100k file), what hardware limits did IBM impose on you for memory and disk space? (For example, you must run on a 4k machine with no more than 32k disk available)?

[garypw...@hotmail.com]

Originally, the DM system was written for a 4K Card/Paper Tape system
(no disk). DM2 added what I believe was a 320K word (640KB) disk.

On Jun 20, 9:49 am, Bob Flanders <bob.fland...@gmail.com> wrote:
> Aside from a user's request (copy a 100k file), what hardware limits did IBM
> impose on you for memory and disk space? (For example, you must run on a 4k
> machine with no more than 32k disk available)?

> On Jun 20, 2011 12:31 PM, <garypwhee...@hotmail.com> wrote:
>
>
>
> > I wrote the original Assembler, some of the I/O subroutines, and some
> > parts of DUP. Later, I managed the team when the macro capability was
> > added to the Assembler, and when RPG was integrated into DM2
>
> > On Jun 20, 8:22 am, Paul Anagnostopoulos <pcanagnostopou...@gmail.com>
> > wrote:
> >> Oh, you have to tell us about the parts you worked on.
>

> >> ~~ Paul- Hide quoted text -
>
> - Show quoted text -

[Richard Stofer]

I believe the internal disk was 512,000 words after formatting. 321
words per sector, 4 sectors per track, 2 tracks per cylinder, 203
cylinders. One word per sector was used for the sector number and 3
cylinders were reserved for defective cylinders. So, net 320 * 4 * 2
* 200 = 512,000 words usable space.

Richard

> > - Show quoted text -- Hide quoted text -

[John R Pierce]

On 06/20/11 9:56 AM, garypw...@hotmail.com wrote:
> Originally, the DM system was written for a 4K Card/Paper Tape system
> (no disk). DM2 added what I believe was a 320K word (640KB) disk.

I thought the card/tape only system wasn't called DM (as in DISK
Monitor). Both DM v1 and v2 supported the 510K word built in disks,
IIRC DM2 added support for the couple external disk options.

--
john r pierce N 37, W 122
santa cruz ca mid-left coast

[Gary]

I think 512K is right. Thanks.

[Gary]

You are right about the name. Duh! How could a diskless system be called a
Disk Monitor system? I'm getting sloppy.

[Paul Anagnostopoulos]

How was the macro facility for the assembler designed? There are some marvelously funky features that were probably designed "to fit." For example, I've never understood why the leading/trailing apostrophe feature was done the way it was.

Even to this day, the IBM 1130 assembler holds a special place in my heart.

~~ Paul

[Gary]

To tell you the truth, I can’t recall the details of the design.  I was in management by the time we added the macro facility, and so I wasn’t involved in a lot of the details.  I do remember however that space constraints continued to be difficult.

 

Gary

 

From: Paul Anagnostopoulos

Sent: Monday, June 20, 2011 4:01 PM

To: ibm...@googlegroups.com

Subject: Re: [IBM1130] Re: IBM 1130 Development

 

[Mark Lossner]

Gary,

You folks did an amazing job with the 1130! I studied the microfiche
of the OS; and, I still write assembler the way you guys taught me.
You were the best I have ever seem. The documentation was plainly
written, with no guessing needed for anything. It is simply amazing
that you did so much with such a resource-limited machine; and, did it
so well that 1130s honestly filled the hearts of their users. Pass the
word to the people who were on those teams. OUTSTANDING!!

Mark


On Jun 20, 10:35 am, "garypwhee...@hotmail.com"

[Paul Anagnostopoulos]

Back in them thar days you had to write compact code. The guys at IBM were masters.

For fun, take a look at the attached listing of the 2-card bootstrap loader. The first card is in 1130 IPL format, so it's a beast. Massive amounts of "clever" code here. I've recently written the equivalent program for my new OS/1130 operating system for the fictitious 1130 model 7. I've attached that listing, too.

~~ Paul

[Peter Diehr]

I never realized that you could use the assembler for this .. I wrote a few boot loaders, but had to work out the machine code, and then punch it myself.

Live and learn!

Best Regards, Peter Diehr

On Thu, Jun 23, 2011 at 1:54 PM, Paul Anagnostopoulos <pcanagno...@gmail.com> wrote:

Back in them thar days you had to write compact code. The guys at IBM were masters.

For fun, take a look at the attached listing of the 2-card bootstrap loader. The first card is in 1130 IPL format, so it's a beast. I've recently written the equivalent program for my new OS/1130 operating system for the fictitious 1130 model 7. I've attached that listing, too.

~~ Paul

[John Doty]

On Jun 23, 2011, at 12:00 PM, Peter Diehr wrote:

I never realized that you could use the assembler for this .. I wrote a few boot loaders, but had to work out the machine code, and then punch it myself.

I never wrote one myself, but a friend once sent me his "card", an IPL format thing that printed his name and telephone number ;-)

John Doty              Noqsi Aerospace, Ltd.

http://www.noqsi.com/

j...@noqsi.com

[Mark Lossner]

Peter,

>I never realized that you could use the assembler for this .. I wrote a few
>boot loaders, but had to work out the machine code, and then punch it
> myself.

I made several IPL cards on the keypunch using the multi-punch ...
then our CE just shamed me into using the assembler. I will never
forget his surprised look when I told him I was making an IPL card. He
just shook his head and proceeded with another patient lecture.

Mark

[garypw...@hotmail.com]

When I was hired by IBM right of college, I was the first programmer
in what would become the 1130 Programming team. Interestingly,
everyone else in the department were former Unit Record marketing reps
with absolutely no stored program computer experience. Prior to
announcement, the 1130 went through at least three significant design
changes, at one point consideration was given to S/360
compatability...think S/360 Model 10. A crucial consideration with
each design was how fast programmed floating point subroutines would
execute. I recoded these multiple times, each time calculating
execution times by hand...no calculators in 1963. When the design as
we know it was finalized, I began work on a bootstrap assembler which
ran on the 1401. We used this to produce code before we had an
engineering model on which to run a native 1130 assembler. Fun times.

> > to the 1130.- Hide quoted text -

[Peter Diehr]

I remember when we were writing the additional floating point routines for the EMU Fortran compiler - we studied your routines in detail, and were very confused at first, and then very impressed.  As I recall, you used the Chebychev expansions as the "best" approximations of the functions.  This results in minimizing the total error over the domain by more closely approximating the range than the truncated series expansions.

Did you work out the coefficients?

Best Regards, Peter Diehr

[Gary]

Actually, the code you saw was written by someone else...a very bright guy.  By the time he was writing the floating point arithmetic, I was heads down writing the Assembler.

 

Gary

 

From: Peter Diehr

Sent: Friday, June 24, 2011 8:06 PM

To: ibm...@googlegroups.com

Subject: Re: [IBM1130] Re: IBM 1130 Development

 

[Tom Watson]

As far as how to calculate functions (SIN, COS, ATAN, EXP, LOG) the
general reference of the day was:
Approximations for Digital Computers written by Hastings. It was
published in 1955 by the Princeton University Press.

On many computers the SQRT function was done by a variation of Newtons
method (test, try, etc...).

On the IBM 1620, it was done a different way called "the odd integer
method", which takes a bit longer to explain, but it works by the fact
that squares can be expressed by a sum of odd integers:
1**2 = 1
2**2 = 1 + 3
3**2 = 1 + 3 + 5
etc.
This works well if multiplication is MUCH more expensive than adding
(which is the case on the IBM 1620).

Funny how mathematics doesn't change much over the course of over 55
years!
--
Tom Watson Generic short signature
t...@johana.com I'm at home now

[Clare Owens]

Ah yes, I still have my Hastings as well as Ralston and Wilf's Mathematical Methods for Digital Computers.  Also hanging on to my TI Transistor Circuit Design book - used that to design the circuits for the Binary Square Rooter 5 of us built as an EE special topics project.  I wrote and ran design analysis programs on the 1620 and plotted the results.  It almost worked but we just didn't understand enough about how to cut down on the coupling of signals between the sections of the device we had built.  But that's another story...

Clare

[Richard Stofer]

Assuming that I could find any of your reference books, is there one
(or more) that might be particularly appropriate for adding floating
point instructions to the IBM1130?


My FPGA version has single cycle signed multiply (25 MHz clock) and
the non-restoring division is fairly quick but if I really wanted to
improve the overall performance, I would think that ramping up the
floating point operations would be the way to go.

I can't even imagine what complications this brings to the software
but, as a first guess, I would just replace the floating point
routines with identically named routines that simply used the new
instructions. As the assembler won't support the new mnemonics I
suppose I would just use a DC statement for the op code.

Any thoughts?

Richard

> > >         > - Show quoted text -- Hide quoted text -

[Peter Diehr]

While I was working at DNA Systems on the compiler turnover, another company was there with an old-style FPGA.  I helped them to test their floating point implementation by linking alternate libraries in the test program, which wrote the floating point answers to a file ... so we had two files, and another program simply ran thru the two files, and noted the entries that didn't match.

There were a lot of differences at first, mostly in the last bit of precision; the engineer made little changes (he basically scraped bits of copper off), and then the tests were rerun.  By  the end of the day, everything matched up!

The alternate library approach was as you described ... 

I don't think I ever knew who the work was being done for; DNA Systems was under contract to help with the testing.  Possibly one of the companies like the GA1830. This would have been spring, 1972.

Best Regards, Peter Diehr

[David N. Lombard]

I used a similar approach to coax microsoft fortran for the 8085 into supporting an amd 9511 math coprocessor. Worked well...

--
dnl
Sent from my Verizon Wireless Phone

[Clare Owens]

Well, both of those books, and the 7040/7044 mathematical routines IBM manual I have all deal with computing the higher level functions, not the actual floating point add, subtract, multiply and divide math so would not be of any help in designing the basic floating point hardware.  They're interesting but really tell you what to do with your FP hardware in order to produce the fastest and/or most accurate results for things like trigonometric, exponential and even linear equation solving.   I never went into actual computer logic design texts other than the British article on how they implemented that binary integer square rooter that we tried to reproduce in the special topics class.  So I'm not any help, I'm sorry to say.

Clare

On Sat, Jun 25, 2011 at 11:40 AM, Richard Stofer <rst...@pacbell.net> wrote:

[James C Field]

Good for you! Figuring out the "wrapped around" pattern of the twelve
column bits on a 5081 to the 16 bit card reader data presentation
exercised my youthful brain and gave me PROFOUND insight into the IBM
mind about keeping old schemes working while introducing new ones.

Jim

[Jeff Jonas]

>>> I never realized that you could use the assembler for this ..
>>> I wrote a few boot loaders,
>>> but had to work out the machine code, and then punch it myself.

Ah, if it's hard for the computer
then it must be hard for me too!

That kinda reminds me of a classmate around 1980
who made an EPROM programmer using just toggle switches
for data & address and programmed his Z80 monitor that way.
I didn't have the patience for that,
preferring front panels for on-the-fly patching and debugging
(having already enjoyed the IBM 1130 for that,
and the PDP11/45 was running Unix full time
so front panels were still in daily use).
I tried making an Altair/Imsai style front panel
and got frustrated with that
so I bought a Timex/Sinclair 1000 and used that as a front panel!

>> I made several IPL cards on the keypunch using the multi-punch ...
>> then our CE just shamed me into using the assembler.

And here we are with microcontrollers with on-chip FLASH
that's programmed directly from the build-environment,
even if it's programmed in "C" or higher level languages!
So the CE was right in directing you to more automated methods!

> Good for you!

Agreed, that's more stamina than I ever had!

> Figuring out the "wrapped around" pattern of the twelve column
> bits on a 5081 to the 16 bit card reader data presentation
> exercised my youthful brain

You mean this format?
http://ferretronix.com/march/computer_cards/1130_plastic_template.jpg

-- jeffj

[Brian Knittel]

When I brought up DMS in the simulator, I had to write a little tool to take an object deck and convert it to the IPL format. (The DMS load deck actually uses both the 1130 and 1800
IPL formats.) (not real decks, of course, card image files). Presumably IBM had something comparable on the 360, where they generated DMS.

Brian

[Eddy Quicksall]

I'll second that. The IBM documentation was very clear and they only wrote what was necessary to give all of the facts. The committees today would easily make a 20 page document 300 pages. I used the IBM 2780 manual and wrote a 2780 emulator for the 1130 (GE bought that). Later I wrote the IBM 3780 and 3741 emulators from their manuals. I remember one thing in particular, the 3780 manual did not say a print line could not span two Bi-Sync blocks so I just did what the manual said and without any assumptions. Later that proved to be a good thing because a competitor made that assumption. Hence the IBM typerwriter division bought my emulator because it would pass the 3780 diagnostic but my competitor's would not.

Eddy

[Eddy Quicksall]

I also wrote one card a bootstrap for the disk. I called it QuickBoot because it seek'd really fast and gave it out as my "calling card". I also punched it using the multi-punch feature (and patched it using left over chips). I got chewed out because I would then duplicate the card and jam up the print heads (I guess I used the 026 at that time where you could not turn off the print heads).

Are you saying the assembler would format and punch the card? What was that feature anyway?

[Eddy Quicksall]

Cool, one of the first things I did in college when we got our 1130 was to write an assembler cross reference program. That was around 1966 or so (I can't remember the years very well anymore).

I remember the first 1130 assembler would punch the binary back into the original cards. That was why everything started in column 21 (I think).

Eddy

[Eddy Quicksall]

Hi Peter, I worked for DNA between 1974 and 1980 and wrote software for DNA between '70 and '73 using the 1130. The GA1830 did not have copper bits, the Digital Scientific Meta-4 did and came out around '70. Dike Summers (a Meta-4 engineer) wrote the firmware including the floating point. Shortly after that Dick Sherin wrote a replacement that required 1 fewer ROM module which DNA marketed. The firmware he wrote gave more precision than IBM's but there was a customer that wanted the same answers as IBM so he added an option to give the reduced precision. Maybe you are thinking of that.

Those copper bits make good conversation. I still have one of those ROM modules. It weighs about 40 pounds and is only 4K. It had 4 boards with copper bits. In those days the word "patch" really had meaning because we scrapped a bit off where we wanted a 0 and pasted a bit, from a roll of bits, where we wanted a 1.

For those that are not familiar with the clones, there were 3 ... the GA 1830, DSC Meta-4 and the CHI 2130. Later, Don Nichols wrote an 1130 emulator to run on the Series 1 and I wrote one to run on the PC. Mine is still in daily use today at Developer Services in Wichita, KS.

Eddy Quicksall

[Peter Diehr]

Hi, Eddy.  We probably met sometime ... I visited a couple of times in the spring of 1971; I don't recall every meeting Don Nichols - most of my interaction was with Bud Blackney.

I recently talked with Dave Morse on the phone; we chatted about old times.  Unlike many of the members of this group, he wasn't much for keepsakes ... 

I wish I had kept one of those copper bits!  It would have made a fun conversation piece.

Best Regards, Peter

[Eddy Quicksall]

There were 3 partners at DNA, Joyce Stout, Don Nichols and Bud Blackney. Dave and I were just grunt programmers (we didn't know we were "engineers" at that time). Don wrote the basic time sharing kernel. I wrote the communications code, memory manager and real time kernel loader. Dave wrote the Fortran compiler (what they now call the EMU compiler), Joyce wrote XDUP and managed the company. Bud wrote the time sharing version of DCPIP and then became the "business man" of the company. He had a massive hart attack around 2005 and passed away ... I really miss him. Tim Mellon (he had his own business) wrote the CLI which was called CYTOS (after the one written at Yale).

I think that system, now running in emulation on a PC, will execute faster than today's systems under Windows. I just had a conversation with DSI and they are running 4 dual active/active systems 24x7 without any crashes ... let's see Windows do that!!!

[Peter Diehr]

Eddie:

Dave Morse and I were co-authors of the EMU Fortran compiler;  I wrote the front end code (parser & syntax, extended variables, and error messages), while Dave did the object code generation, and most of the new library routines.  It took us about a year to develop a complete product.    I gave talks on the EMU Fortran compiler at the December 1970 COMMON meeting in LA, and again in April 1971 in Chicago.   EMU distributed the compiler for a nominal charge (via tape) in object form, or in source form.

DNA Systems purchased a source code tape from EMU, and then provided support for commercial clients.  I did the original modifications to the compiler to interface with CYTOS ... one evening, up in Saginaw.   I think I spent a total of two weeks up there, doing various tasks for the turnover.  Bud paid me $2,000 for my work.

Bud offered to hire both Dave and myself, but I had just started a graduate degree (MBA), and declined.   I met Joyce Stout and Tim Mellon a couple of times.  I may have met you as well, but I don't recall.  

I last spoke with Bud a few times in 1984-1985, after moving back from San Francisco to Ann Arbor.  I had been out of touch since 1973, but DNA Systems was still around, though the IBM 1130 was long gone.   It remains my favorite computer ... must be first love!  

I don't program much anymore ... too busy designing novel electron guns, and doing work with carbon nanotubes.  Physics makes progress at least as fast as computers (often drives the process), but the old knowledge never becomes obsolete,  You just have to refine bits of old knowledge a little bit to make it remain useful - so Newton's Laws  are still good today as long as you understand where and when the quantum and relativistic regimes need to be considered.

Best Regards, Peter Diehr

[John R Pierce]

IIRC, Cytos used memory swapping on the 2130, where the hardware had 64K
core, the top 32K was used for the CYTOS OS, and each user process was
swapped into the lower 32K, so time slicing was rarely more than once
per second. If 6 or 8 people were active, it could get really slow.

if folks have this running on an emulator, it seems like it would be a
pretty easy hack to add, say, 16 32K 'normal' memory banks and patch the
swapping code to simply select a bank and voila, process swapping would
be instantaneous.

--
john r pierce N 37, W 122
santa cruz ca mid-left coast

[Eddy Quicksall]

Tim Mellon did write a version of CYTOS that had time sharing capability and it would run on a native 1130. I know one was on a Meta-4 (I think at Lawrence Tech) and one may have been on a 2130 (I can't remember). Generally though, the 2130 people bought TSO which had CYTOS as a part of it ... some people may actually remember TSO as "CYTOS" because that was the front end. The last I talked to Tim he didn't like to use the internet nor email ... I haven't talked to him in years so now would be a good time to checkin and I'll tell him about this forum ... I think he would love to chime in.

TSO began as a disk swapping system then Don Nichols added drum swapping then memory swapping then memory mapping. With the memory mapping Don overlapped disk I/O with a "task switch" which made it even faster. I wrote the memory manager to give it an option for both memory mapped and/or disk swapping. The current system running running at DSI is strictly memory mapped. I originally wrote the disk and printer routines so they would make use of the overlap but after the PC got really fast it didn't make much difference and it was not easy to write asynchronous code for the laser printer and large disk. But I did overlap the tape I/O with task switches. With TSO/CYTOS now running on a PC at 3.5Gig or so and memory mapping, everyone is in L2 cache and, as far as I know, they barely know anyone else is on the machine.

I'm still into programming. I now work for Cisco where I work on their Nexus 7000 ... a very big and capable switch.

I have always had a love for physics. Especially hearing about high energy physics (I don't understand most of what I hear but it sounds cool).

I just noticed you are in Santa Cruz. I'm in Santa Clara. We should try to get together sometime. Send me an email at my_first_name at my_last_name and the "com" extension.

[John R Pierce]

On 07/27/11 7:11 PM, Eddy Quicksall wrote:
> Tim Mellon did write a version of CYTOS that had time sharing
> capability and it would run on a native 1130. I know one was on a
> Meta-4 (I think at Lawrence Tech) and one may have been on a 2130 (I
> can't remember). Generally though, the 2130 people bought TSO which
> had CYTOS as a part of it ... some people may actually remember TSO as
> "CYTOS" because that was the front end. The last I talked to Tim he
> didn't like to use the internet nor email ... I haven't talked to him
> in years so now would be a good time to checkin and I'll tell him
> about this forum ... I think he would love to chime in.

yeah, it was TSO/CYTOS, sorry. old neurons. I only worked for CHI
for about 6 months in 1975, then never saw another 2130 again. and,
afaik, that was the only time I saw TSO/CYTOS.. Shortly after that I
became involved with early microcomputers and in 1977 ended up at DRI
(Digital Research) working on CP/M.

[Eddy Quicksall]

Did you know John Curry? He worked at CHI around that period before he came to DNA. I forgot the owner's name but he wanted me to come to work for him. I turned him down because I was really good friends with the DNA bunch. Did you know Ken Borassa from Corpus Cristy TX? He was a CHI bigot until around '94. I kept track of him for a while but lost track just around that time.

[John R Pierce]

the name John Curry rings a very faint bell, but I can't place it. As I
said, it was 6 months in 1975. there were 3 of us in the software
group, I was low man on the totem pole. Sheesh, that was 35 years ago,
hah.

[da...@i-f-s.com]

I used CYTOS on an DSC Meta-4 at University of New Haven during the mid seventies and up through around late 1978, when I worked as a programmer/analyst there. I think we also had the DNA Fortran compiler too because the university did all of their administrative and student reporting systems in Fortran. As I recall, the DNA compiler had many extensions to make "business" programming much easier. I met Tim Mellon there while he was working to install CYTOS.  It was pretty slick and turned a system that only did batch jobs to a real-time multi-user system.

I did mostly Fortran programming then and also did a bit of RPG, COBOL and ASM. Wrote an ASM overlapped I/O card listing program called QLIST.

[Eddy Quicksall]

I worked for DNA during that period. I got to know Tim very well. He sold CYTOS to DNA but kept a few customers and I think University of New Haven was one of them. Yea, I think they used our Fortran compiler. I'm currently doing some consulting for a company that still uses TSO/CYTOS on an emulator I wrote.

[David Best]

I knew Don, Bud and Joyce very well.   I joined General Automation in 1969 and ported DM2 to the GA 1830 (which required rewriting all of the I/O drivers, and modification to the Core Load Builder to remove direct references to memory cells 1,2,3 which on the 1130 were index registers, but were not on the 1830).   I got to know the DNA folk before joining GA as a customer of their XDUP when I was lead programmer at Aurora Metal Systems in Illinois.   I got GA to hire Don to port the 1800 TMS OS to the 1830 and we had some great times tearing up LA on the weekends together.   When GA went into religious schism at the Board of Directors level, I joined DSC and did all the pre- and post-sales tech support for the Meta-4, and got to know Tim Mellon at that time.   When DSC got into financial trouble I jointed DEC and lead the development of the 2nd generation VAX/VMS system (Vax-11/750) before going to Intel.   I still dream about 1130 assembler code and miss the people a lot.   Are Don and Joyce still around - anyone know how I might reach either of them?

David Best

[Eddy Quicksall]

Did you know Berry Kennedy at GA? He designed the DSM interface and changes necessary for TSO. The last time I talked to him in the mid 90’s he was working at Intel. We had a nice visit when I was in CA once. How about Kim Crosser at DSC? I worked mostly with him in the software area.

 

Don, Joyce, Dave and I are still around. Bud passed away in the late 90’s … I really miss him. Don quit firmware in the early 80’s (what a waste). I saw Joyce and her husband last year … we had a really nice visit. Dave moved on to GM in the early 80’s.. Richard J. Sherin (who wrote our IBM 1130 emulator for the DSC Meta-4) quit programming in the mid 70’s and went into design and manufacture of a Lasik blade. Tim Mellon quit programming after writing CYTOS. We are all retired now.

 

Bud, Joyce and I shot on pistol teams (opposite teams so it was sometimes dicey). We were all into flying too … DNA had a Cessna 210, a Cessna 182 and a Cessna 150. Those were the days!

 

If you liked TSO then download my system and have fun. It is here (Dropbox keeps changing the URL every time I update it): https://www.dropbox.com/s/rn246a4w7pqiini/Distrobution.rar?dl=0

 

Eddy Quicksall

My personal email address is <my_first_name> @ <my_last_name> .com

 

From: ibm...@googlegroups.com [mailto:ibm...@googlegroups.com] On Behalf Of David Best
Sent: Wednesday, August 24, 2016 4:29 AM
To: IBM1130
Subject: Re: [IBM1130] Re: IBM 1130 Development

 

I knew Don, Bud and Joyce very well.   I joined General Automation in 1969 and ported DM2 to the GA 1830 (which required rewriting all of the I/O drivers, and modification to the Core Load Builder to remove direct references to memory cells 1,2,3 which on the 1130 were index registers, but were not on the 1830).   I got to know the DNA folk before joining GA as a customer of their XDUP when I was lead programmer at Aurora Metal Systems in Illinois.   I got GA to hire Don to port the 1800 TMS OS to the 1830 and we had some great times tearing up LA on the weekends together.   When GA went into religious schism at the Board of Directors level, I joined DSC and did all the pre- and post-sales tech support for the Meta-4, and got to know Tim Mellon at that time.   When DSC got into financial trouble I jointed DEC and lead the development of the 2nd generation VAX/VMS system (Vax-11/750) before going to Intel.   I still dream about 1130 assembler code and miss the people a lot.   Are Don and Joyce still around - anyone know how I might reach either of them?

--
You received this message because you are subscribed to the Google Groups "IBM1130" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ibm1130+u...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Eddy Quicksall]

Up until I wrote the emulator in the mid 90’s one of our DNA customer’s was still using 2 interconnected GA 1830’s and one spare. I replaced them with my emulator and he is still using it today for his service burrow.

 

Eddy

[Eddy Quicksall]

Dave, you  may be interested in FMTR.S(0D02). It is the formatter for the GA 300MB disk (we thought that was big in those days). Don wrote it.

[John R Pierce]

On 8/24/2016 4:53 PM, Eddy Quicksall wrote:

Dave, you  may be interested in FMTR.S(0D02). It is the formatter for the GA 300MB disk (we thought that was big in those days). Don wrote it.

300MB *was* huge in the 1970s.   The CDC 9766 SMD drive circa 1975 was the first widely available 300MB drive, each drive was the size of a washing machine, top loader with the 12 platter 14" disk pack.   These were based on the technology of the IBM 3330 but at a higher density (3330 model 1 was 100MB, model II was 200MB).   DEC sold the CDC 9766's as RM05.

Those SMD drives transfered raw data at the blinding rate of about 1Mbyte/second.

I'm remembering CHI 2130's used a 10MB or 20MB 12 platter disk in various configurations, these were about the same physical size as those 300MB CDC drives, but considerably cheaper.  I think they were based on IBM 2314 packs, but we were using Memorex (660?) and other drives.

-- 
john r pierce, recycling bits in santa cruz

[Eddy Quicksall]

In ’74 at InsuraCalc in Miami, we had a Calcomp “2314” on our Meta-4.

 

Do these attachments look familiar? They were still in use when I replace the 1830’s with my emulator around 1995. For non GA familiar people, the 1830 (a cross between the IBM1130 and IBM1800) was a clone of the 1130 with enhancements to run TSO. There are 3 1830’s in these  pictures and two act as failover for each other. The device in the middle is a tape drive. Dave will probably be able to say how they were connected electronically. In my emulator I connected the PCs via a serial line. Each computer could share the drives. On the CHI 2130 we had shared memory.

 

The ROM board was for the Meta-4. To make a patch, we would scrape off a bit to get a 0 and paste a bit onto a blank spot to get a 1. Believe it or not, spare bits came on a roll similar to a roll of chips today. Each bit had copper on front and were sticky on the back. That is when “firmware” really had meaning. When I wrote BIOS at AMI I kept some of this ROM on my desk where I would use it to show to the new comers.

 

We were really advanced in those days.

 

Eddy

 

 

From: ibm...@googlegroups.com [mailto:ibm...@googlegroups.com] On Behalf Of John R Pierce
Sent: Wednesday, August 24, 2016 8:17 PM
To: ibm...@googlegroups.com
Subject: Re: [IBM1130] Re: IBM 1130 Development

 

On 8/24/2016 4:53 PM, Eddy Quicksall wrote:

--