All Terrain Thinking

A Compendium of things I think are Important

Historical Perspective on Today's' PC

Memories of Burroughs

These (edited) memories have been contributed by visitors to this web site.

 

Subject: Burroughs B17xx Series of computers - From: Gary 'ROD' Rodriguez, Systems Architect & CEO

Thanks ever so much for your efforts to capture and document the Burroughs B17xx Series of computers.

I had the distinct pleasure of working on these machines as a Field Engineer from 1974 thru 1978, or thereabouts. It taught me just a whole lot about abstract computer architecture, especially since it blurred the boundaries between those functions performed by hardware and those by software.

Since the B1700 had a vertical microinstruction, with a register-to-register definition of traffic accross the Main EXchange bus, it would have to be considered to be a prototypical instance of RISC architectures. The B1800 series introduced the equivalent of 'nanoinstructions' which would be called Very Long Instruction Word (VLIW) in contemporary parlance. We also started to see a creeping of TTL components replacing the CTL logic which was my most significant disagreement with the B1700 implementation.

The B1700 main memory was Intel 1024x1 DRAM which was built to Burroughs specs. Since the big B always did things on the cheap, the parts were packaged to Burroughs specifications. As a result, the DRAM's hermetic seal would leak ambient air into the device packages. Computers would often be shut down over the weekend (but not the air conditioning) and after awhile, sufficient humidity (~40% in most computer rooms) would condense on the die -- and on some Monday morning when the systems were powered up, numerous hard memory failures would occur. This unfortunate situation ruined the reliability record for the system.

The 'cache' memory for the system was broken down into boards which had a capacity of 4K words, and if my memory serves me right this should represent a 4KByte capacity of TTL SRAM per partition. Up to 4 of these could be had in any given system, and the MCP or a language S-code interpreter would be staged in each partition. This meant that there could be FORTRAN, COBOL, ALGOL or MIL (MicroInstruction Language [assembler, no doubt]) interpreters all running concurrently.

Something else that should be understood about these systems which could run circles around a large part of the Burroughs product line, short of the B6800. Burroughs deliberately slowed them down so that they would not threaten their installed base of B2700, B3700 and B4700 relics. Several methods were used, and one of those was to constrain the bandwidth of the multiport memory controller. Another, obvious one, was to retard the base system clock rate.

In my stay with Burroughs I developed a reputation for being able to debug memory and logic failures to the board and frequently to the chip -- from the front panel. There were few locations of the system which were not mapped into my brain. I really, really miss these machines.

The Reed-Solomon encoding of the 16bit instruction was also somewhat remarkable at the time.

Unfortunately the Wikipedia discussion of the B1700 needs considerable cleanup. I may have to take that on, in my copious spare time, sure.

I've never really had anyone to share this with before.

Thanks,

Gary 'ROD' Rodriguez

 

Subject: Series L/TC - From: Henry Griggs of Virginia, USA

Wow. All those photos! They really brought back the memories of the Burroughs gear. I liked the L series of machines, but the B80 was my favourite. I did a lot of work on those things, but only the really cut down ones, not the really fast ones.

I've got a bit of Burroughs memorabilia stuffed away - the little yellow Assembler code notebook, notepads with Burroughs logos, some advertising material...  I'll scan what I've got and send the images to you so you can enjoy them too. I might scan that entire little assembler coding book. I spent so much time with it, it was years before I could forgot those codes!

 

Subject: B700 - From: Bill Roberts of New York State, USA

I worked for Burroughs in Downingtown, PA, from 1972 to 1977. We made the B700 and B800's there. I wrote the micro-code that allowed the B700 to do addition and subtraction (in decimal so there were no binary rounding errors). I also headed up a team to add up to four key-to-disk microprocessor-controlled (B7* microprocessor) keyboards to allow key-to-disk data collection and verification while the B700 continued it's normal processing. 

The key-to-disk system was the AE500 system. This system was being developed with all the following going on at once: modification of the disk control hardware to allow control by both computers (B700 and B7*) at the same time; development of the Assembler to do the micro code for the B7*; development of an interpreter to read the audit entry instructions; development of a compiler to compile the AUDit EntRY (ADURY, the official name of the compiler) instructions to something that could be interpreted; and modifications to the micro code of the B700 to allow for multiple disk accesses.  The only debug tool was an HP logic analyzer with 8 probes!

I remember hooking up the TD700's (terminal display units) to the B700. 

The B700 was originally released with RPG-III. Although I wasn't involved with the development of that system, I did decide to learn the language, and wrote a program to keep track of my personal finances for (US) income taxes. This helped to uncover some early bugs in the compiler, and, as a result, my personal finances ended up as a part of the test suite for the B700! I at least got to change some of the numbers before it got out of the plant and on to our sales forces in Detroit.

We intended the B700 to compete with the IBM System 3 (also an RPG-III machine), with the B1700 as the "big brother" for when the customer needed a faster machine. As a result, we were one of the first purchasers of the IBM System 3, a fact IBM made use of in its sales promotions. Well, we needed something to run our tests on! In both compile time and program run speed, we were able to beat the System 3. But the real problem was that we totally wiped out the B1700 as well! Turns out the B1700 disk-swapping system swapped out the disk controller at every opportunity, and that had a negative effect on performance. Burroughs management was not amused, several engineers at the B1700 plant did not get their Christmas bonuses that year, and there was a lot of hard feeling between plants for a while.

We took pride in creating demo's that really did stress the machines as much as real life applications. But remember, we were the developers, and quite often did not realize how the machines would really be used.

Finally, I did the design on the B800 job-swapping system. I left before that project came to fruition, but it was a fun project. If you ever played the "Battleship" game on the B700, I did the firing algorithm. It was quite a machine, and (IMHO) well ahead of it's time.

 

Subject: Series B80 - From: Joe A. of Location unknown - 3/2008

The original B80 had 80kb of memory and no floppy drives. It used an FSK tape cassette drive to load programs and save data. The only output was the 132 character printer.

All this in a very neat package. I am pretty sure it was introduced around 1976. I do recall a review that indicated that this was the small business owners dream machine because of it's cost. You could by all this great technology including the business software for just under $100,000. If I recall right the memory access time was 2.5 micro seconds, even slow compared to other Burroughs products of the day.

I worked on the project in the EMSO plant. This plant was responsible for producing the memory boards for all of Burroughs computers (mostly mainframes).

 

Subject: Reply to: Joe A. From Ed - 3/2008

It is always comforting to find another person with memories BUPC, before ubiquitous PC's. In many ways I feel that what we have become complacent to, computing power in phones, watches, games and toys are in fact not improving our society, but rather destroying the civilization that we so proudly built, yes using technology, but not so much a slave to it.

""Back in the Burroughs"" days it seemed like computers were a logical tool to do more with greater accuracy. Today they seem to have raised the bar to a point where the bare minimum requires perfection....

 

Subject: Series B80 - From: Joe A. (continued) - 3/2008

I have to agree with you Ed.  Society has become very dependent on tech, especially, PC's, Cell phones and cable/satellite TV. 

It's very interesting to note that each has spawned it's own cost recurring industry.  Most people pay monthly fees for each and I am sure it takes a pretty good bite out of their monthly income.  The PC can be used for so many great things but we also have people out there using it as a tool to molest our children.  As people continue to replace God with Tech our demise of this civilization as we know it is certain.  It's not "if" but "when". 

Back when I was working for Burroughs, I used a machine call the ADAR to test computer memory.  I designed the interface and developed the programs  for memory testing using this device. At that time the government would not allow the ADAR out of the USA because it was feared it could be used as some sort of guidance system for nuclear bombs.  Mainly since you could map an X and Y coordinate it was feared that could be somehow converted to mapping enemy targets.  Boy things sure have changed.  Now we go out of our way to send our high tech military stuff to be made by China and other foreign countries.  It makes one wonder (now that almost all of the micro-chip tech is made outside the US) if one of these countries will include an "OFF" command inside the chip that they would be able to shut down our infrastructure as well as our weapons when ever the need arises.

It was good to see someone keep this early computing (Burroughs history) alive.  It was an exciting time for me.  I got to work on a brand new technology, solid state memory, when it first became available around 1973.  I had programs that found problems with these chips that no one else had.  Burroughs use to trade my programs for better pricing and first dibs (since production yields at that time where very low).  Being only 20 yrs old I had know idea of "consulting".  Instead, Burroughs owned the rights to everything I did for them.  However, by 1976 all the chip companies in Silicon Valley knew me.  I got an offer for triple my salary and a free move (which I declined, I met my future wife that same year).  Three years latter I left the industry to work in a new tech closer to where I was living, telecommunications.  I had a lot of fun in those days.  Joe

 

 

If you have anything to add to these pages: Send mail to webster[at]emcarthur[dot]com using the subject "Borroughs Hardware", with questions or comments about this web site.

Last modified:30 March 2008

Add to Your Social Bookmarks: Del.icio.us Digg Furl Reddit Simpy Spurl Y! MyWeb - Netscape - StumbleUpon StumbleUpon