These historical notes consist of E-mail from various people who played parts in the development of the PDP-8 family or who were early users.
Date: Sat, 23 Mar 1996 14:01:06 -0500
In the way of introduction, I was the PDP-8 Software Development manager in the late 1960's during the time when PS/8, later called OS/8, was created. I started working for DEC in the summer of 1967. At that time, Roger Pyle was the PDP-8 Software Development supervisor. Roger, another person whose name was John, and one or two others whose names elude me right now, and I, developed the 4K Disk Monitor system and its associated software. At some point, Roger moved on to another project, and I became the PDP-8 Software Development manager (or supervisor, not sure which.) I moved into PDP-8 Marketing (I think it was in 1970) and then Dennis (Denny) Pavlock became the PDP-8 Software Development manager for several more years. This was a challenging and interesting period; and the development of PDP-8 software remains, in my opinion, unique in the history of computing.
The reason for its uniqueness is that all of Digital's PDP-8 system software was developed by an unusually small group of creative, highly motivated and productive people in an environment where there were no standards and very few rules. It was also unique because, as I'm sure you know, PDP-8 code had to be constructed out of subroutines that consisted of no more than 128 12-bit words, and programs had to be able to fit in 4K words of memory, so code had to be extremely compact.
Prior to about 1969, almost all DEC software development for the PDP-8 family of computers was done using a cross assembler. Initially that was done on a PDP-6 computer, and I think the cross-assembler was called PAL6. Some programmers would write their programs on coding sheets, give those pages to the "Tape Prep" group, which consisted of a group of typists who would type the code, typically on a PDP-5 computer with an ASR-33 Teletype using the editor. The Tape Prep person, when finished, would output the program on the "high-speed" paper tape punch and then return the coding sheets and tape to the programmer. The programmer would then load that tape into the PDP-6, run PAL6, copy the binary file to the high speed paper tape punch, and save the program source on a DECtape. We would then load the binary paper tape on one of the development PDP-8 computers and try to run the program.
Of course it was likely that the program would have assembly errors, so the programmer would edit the code using TECO, and try again, from time to time saving the latest version on DECtape. The PDP-6 was not very reliable then, maybe on a good day only "crashing" a few times. So moans and groans would periodically be heard coming from the PDP-6 computer room each time the system crashed, as programmers thought about how much editing they would have to do over again. Also in those days, disk storage was extremely limited. The PDP-6 that we used in the Maynard mill had a drum memory with a capacity that would have looked small on a PC in the early 1980's. So DECtapes would constantly be spinning on the front of the PDP-6 as programmers copied their code to or from the tapes. The PDP-6, by the way, was DEC's 36 bit computer and was roughly equivalent to a mainframe that supported time-sharing.
Once PS/8 became useable, assembly language software development on the PDP-8 became common. Prior to the development of PAL8, the symbol table in PAL III was too limited to assemble anything other than small and medium size programs. The MACRO-8 symbol table was even smaller. PAL8, however, was derived from the MACRO-8 paper tape assembler (maybe a descendent of MACRO-8.) The macro features were removed, and the symbol table was moved to extended memory, vastly expanding its size. But PAL8 was still very slow until someone, I think Richie Lary, implemented a binary search on the PAL8 symbol table. That remains, in my mind, a defining moment, where the PDP-8 became a useful computer rather than an interesting toy. Subsequently, OS/8 and all associated system software and DEC PDP-8 programming languages could be developed entirely on the PDP-8 using OS/8 and the PAL8 assembler.
There are a few names that could be added to the PDP-8 Who's Who. Roger Pyle deserves credit for much of the design and development of the 4K Disk Monitor system. Lary Portner, who later became the Digital Software Engineering manager and a DEC VP, wrote the 4K FORTRAN compiler (although he confided in me once that he was not particularly proud of that.) Henry Burkhardt (sp?), who went with Edson deCastro to start Data General, wrote the 4K FORTRAN operating system. Even though it was far short of real FORTRAN, it was the first high level language on the PDP-8. The person mentioned along with Richard Lary and Ed Friedman, was Paul Kneuven.
Dick Palmer (no relation to Robert Palmer as far as I know) also worked on the PS/8 software. Dennis Pavlock should certainly receive credit for leading PDP-8 software development in the early to mid-1970's. Richard Lary, now (I think) a DEC Senior Consulting Engineer, also deserves credit for his work on COS/DIBOL software, particularly for the development of a superb Sort/Merge program and for his role, along with other in the development of OS/8 FORTRAN-IV. The DIBOL language was defined by a person whose name was, I think, John Cohen. The significant feature of DIBOL was that it supported decimal arithmetic, which would allow precise math vs. most other PDP-8 software which at that time was limited to 6 digits precision. The problem with the early version of DIBOL was that it was based on the DECtape monitor system and its required use of paper tape was incredibly cumbersome. DIBOL became viable when the COS operating system was developed, but I don't think it was widely accepted because it was not a standard language. I wish I could remember the names of some of the other people who worked on COS besides Richard Lary. Steve Wellcome, who also did some OS/8 work is the only name that comes to mind now.
There were many other significant PDP-8 software development efforts both inside and outside of Digital that truly deserve credit. The WPS Word Processing System was used widely within Digital and was sold to customers during most of the 1980's. There was a major PDP-8 based typesetting effort within Digital.
A significant PDP-8 operating system and FORTRAN compiler was developed at, I think, the University of North Carolina, and some of the inspiration behind PS/8 came from a programming system that was developed at the Univ. of Michigan. One of the people there was Kurt Metzger. It is true that while Richard Lary was working through the night developing what later became PS/8, DEC management thought we were modifying the U. of Michigan programming system. This was done because the PDP-8 Product Line managers had indicated that they did NOT want to spend the money to develop a new operating system from scratch. They were, however, willing to support modification of an existing system.
An interesting program COLPAC, roughly a cross between FOCAL and BASIC, (if my memory is correct) was developed at Carlton College in Canada. [no, in Minnesota, see Jon Dreyer's note] Bob Hassinger, an active DECUS member for many years, did extensive PDP-8 software development and was one of a small number of valued test sites. I wish I could remember more names, because I worked with some wonderful people during my PDP-8 software years. Some of those names will probably come to mind if I think about this more.
Date: Mon, 25 Mar 1996 09:27:59 -0500
A few more names came to mind as I was thinking about this over the weekend. Chuck McComas was another programmer who did quite a bit of PDP-8 system programming along with Ed Friedman, Paul Kneuven and Dick Palmer. Bill Long was the Senior Vice President who had overall responsibility for the PDP-8 during the late 1960's and early 1970's. Nick Mazareese was the PDP-8 Product Line manager during the late 1960's and was known, within the software group, for apparently believing that software was a necessary evil and that as little should be spent is this area as possible. Boy how I would love to have seen a debate between Nick and Bill Gates!
Mike Ford was, I think, a PDP-8 marketing manager in the mid-1960's who left DEC to start a business selling small business applications based on the PDP-8. Unfortunately, I don't think things went too well for Mike. Hopefully he didn't sell all of his DEC shares at that time. He was only one of many people who tried to apply the minicomputer of the 60's and early 70's to meet the needs of small business, but largely failed because the cost was too high and the hardware and software was just not adequate at that time.
There was one PDP-8 hardware engineer in particular who deserves recog- nition, and I think this was Don White. He was responsible for a lot of the engineering behind both the CPU development and some of the PDP-8 peripherals. He was also heavily involved in development of the PDP-8A and the later DECmates, etc.
There was another software manager between the time that I left the programming group and Denny Pavlock took over. I've been trying to remember his name, but so far it has not come back to me.
There were lots of interesting and funny stories about Richie Lary going back to that period, but I've run out of time. I will try to get back to you
Date: Mon, 25 Mar 1996 09:27:59 -0500
No history of the PDP-8 would be complete without a mention of George Thissell. George took over management of PDP-8 software after I left the group, and as best as I can recall, Denny Pavlock worked for George at that time.
It was under George's management, that the PDP-8 OS/8 FORTRAN-IV was developed. Even though it required floating point hardware, the development of such a compiler, run-time library, and operating environment on the PDP-8 with its memory and address limitations, and with the tools that were available at that time was an incredible accomplishment.
Date: Mon, 17 Mar 1997 23:03:51 -0500
COLPAC was in use by 1972 at Carleton College in Northfield Minnesota (www.carleton.edu). I believe it was also written at Carleton; if I remember the COL stood for "Carleton On-Line"). From what I remember, syntactically COLPAC was very much like FOCAL, from which it was derived, but it had some builtin functions to handle graphics output to a Tektronix (I think) storage scope as well as joystick input. We had four 8k PDP-8/L machines that we could run COLPAC on. These were essentially personal computers, and those of us who learned to program them realized what an opportunity it was to have a whole machine to ourselves in those days.
Carleton also made heavy use of a PDP-8/I TSS-8 system. Our hardware whiz, Randy Riegstad, hooked up a PDP-8/L to it to buffer i/o between the 8/I and two of the disks, so we sort of had an asymmetric multiprocessor.
Carleton was also the source of FOCARL, another FOCAL dialect that had a lot of enhancements that unfortunately I don't remember.
Carl Henry might remember some of the details. Dean Krafft might also remember some more of this.
Date: April 6, 2004 12:18:03 PM CDT
A few additions:
Stan Rabinowitz worked with Richie and me on PDP-8 DIBOL. We did it around 1972. I think Stan did the compiler and some other utilities. Richie did the interpreter (the compiler produced meta-code that was interpreted) and the memorable sort program, and I did the more-or-less operating system and some other odds and ends. In retrospect, we probably should have used OS/8 as the operating system, but it seemed to make sense at the time to do our own.
The three of us were on 12-3 in the mill. I shared an office with Richie, which was an Experience, and Stan's office was next to it, with Herb Jacobs. At some point Richie decided to paint our office, so he got some paint and painted the plywood partitions. Not to be outdone, Stan and Herb completely covered the walls of their office with contact paper and put down a cheap shag rug on the floor. The rug, of course, was soon ruined by the oil that seeped out of the wood floor...not that it was much of a rug to begin with.
Date: December 27, 2004 1:20:08 AM CDT
I was browsing the net the other day reading up about DECtape when I came across your site. I was reading the page entitled "What operating systems were written for the PDP-8", and right down the bottom was a paragraph:
AMOS, an operating system for the PDP-8/E with TD8E DECtape interface, was a very small system developed in Australia or New Zealand and supporting assembly and text editing on a 4K machine.
Well I was surprised to read this as I and a friend wrote this sort-of OS back in 1973 or 1974 when we are still at high school here in Christchurch, New Zealand. We were about 15 years old when we started playing about on the school's PDP-8/E which had an 4K core memory and an ASR33 only but later they purchased a DECtape unit but couldn't afford the extra 4K to run OS8 so we wrote this little MOS system to fit in 4K. The only thing is that we called it MOSS (Magtape Operating System Software) so I assume it's the same thing as you called AMOS in your summary.
We did submit it to DECUS (#8-770 from memory), but I have never heard if anyone ever used it or anything else about it until stumbling across your website. Like many who used DEC systems I still have a great fondness for the company and its products, and I'm really pleased that people such as yourself keep the memory alive. I could kick myself for not keeping the old PDP-8 stuff I had, all I've got now is a copy of Introduction to Programming that I learnt PAL-III from (my first language) and a whole lot of PDP-11 stuff. I guess once you get past 40 you start looking back to the things that affected you when you were young, and the PDP-8 certainly had a great influence on my life (I'm a programmer). Anyway just for the record my name is Simon Young and my fellow author of the PDP-8 AMOS [MOSS] system was Ben Lewis. We were probably 16 or 17 when we wrote it.