A Short History of CP/M-86
The Early Battle for 16 bits
Kathryn Strutynski was born on the 5th of February in 1931 in Nephi, Utah. Strutynski attended Brigham Young University from which she attained a degree in mathematics. Her first job after leaving college in 1953 was teaching high school math. Strutynski then moved to California and worked for Pan Am. This was short lived as Pan Am moved their offices, and Strutynski then worked at McGraw-Hill, and then at Bechtel. Her time at Bechtel was the most impactful of her early jobs from what I can tell. The company purchased an IBM mainframe and sent her to all of the training courses offered by IBM. This resulted in her leading a team that developed and maintained a database system on the computer that the company then used for more than a decade. It was also during this time that she met and married Alfred Waldemar Strutynski. After their marriage, the couple moved to Monterey, and Mrs. Strutynski gained employment at the Naval Postgraduate School where she also began taking classes and working toward her master’s degree. She met Gary Kildall there and the two became friends.
Strutynski became the fourth employee of Digital Research Incorporated on the 1st of February in 1979. Kildall had walked into her office at NPS at the end of the workday, picked up her briefcase, walked out to her car with her, and then told her he could afford to hire her. She accepted, and it seems that Kildall was quite lucky that she did. Strutynski did the bulk of the work on porting CP/M to the SoftCard, and then did some work on CP/M versions 2, 2.2, and 3.
While Digital Research may have been a little slow to adapt their operating system for new processors, they were working on it. In 1980, Strutynski was put in charge of DRI’s OS efforts on 16 bit CPUs, and before the IBM Personal Computer launched, Strutynski and her team had ported CP/M to the Intel 8086/8088, and they’d sold licenses of CP/M-86 to AST and Altos. These were followed by Piiceon, Artelonics, and Sirius Systems.
In November of 1981, XLT86 was available to convert 8080 assembly to 8086 and could be run on CP/M, MP/M, or VMS. For people running CP/M or MP/M the price was $150 (about $515 in 2024) and the VAX version cost $8000. Around this time, Digital Research managed to get IBM as a license holder of CP/M-86 for the IBM Displaywriter word processor.
MP/M-86 version 2.0 (despite being the first release for 8086) was shipped in November of 1981, and Digital Research referred to it as the most powerful operating system to be made available for the 16-bit microcomputer. Specifically, they noted that it’s queue system and multi-tasking were state-of-the-art, and they noted that it had networking capabilities.
By the end of 1981, around twenty five percent of 8088/8086 installations were using CP/M-86. This isn’t at all surprising considering the CP/M-80 had been the 8 bit standard with more than four hundred different manufacturers shipping CP/M-80 machines. It helped that Digital Research was well established, and their customers who’d shipped 8 bit CP/M machines could turn to them for 16 bit operating systems as well. The future looked bright for DRI, and many expected that CP/M-86 would win the 16 bit market.
In 1982, Digital Research released CP/M-68K for the Motorola 68000. This was initially written in Pascal following DRI’s acquisition of the MicroSYSTEMS company in Solana Beach. This port was originally made for the Atari ST, but Atari chose to go a different way. Fortunately, the effort wasn’t wasted, and this port shipped on the Sord M 68 in 1982. This was a dual architecture machine featuring both a Zilog Z80 clocked at 4MHz and a Motorola 68000 clocked at 10MHz, and it shipped with up to 1 MB of RAM.
May of 1982 saw the release of the DEC Rainbow 100. This machine shipped with a Zilog Z80 clocked at 4MHz, an Intel 8088 clocked at 4.815MHz, with up to 896K RAM, up to four 400K floppy disk drives, optional winchester disk drives, and serial and parallel ports. This machine was bizarre for a number of reasons. The dual slot floppy drives required that the top slot disk be inserted upside-down as the head assembly was in the middle. The machine also made use of a proprietary disk format, so not just any disk would work. This machine wasn’t expandable, and those cards that could be upgraded used proprietary connectors. Then, there’s the two CPUs. Running CP/M, if an application was written for 8 bit it would run on the Z80, and if it were written on 16 bit the application would run on the 8088. CP/M handled the CPU assignment. For standard operation, the Z80 handled the floppy disk drives while the 8088 handled video, keyboard, printer, and any optional hardware. There were other systems featuring both the 8 bit and 16 bit CPUs of that time, but the DEC Rainbow is certainly a highlight for its quirks.
All of these ports were based upon CP/M and MP/M versions from the 8 bit series. In particular, CP/M 2.2 with some enhancements from MP/M and CP/M version 3. Yet, there were some other versions which make all of this more confusing than it already might be. Were one to have purchased a computer from ACT, Olympia, or a handful of others manufacturers, there was CP/M-86 Plus. This appears to have been an OEM only offering from what I have been able to find. Also, I did find that version 3.1 of CP/M-86 Plus was released in October of 1983. As for what it was, it appears to have been CP/M-86 with DESPOOL (a multitasked printer spooler developed by Strutynski), polled interrupt drivers, and clock tick support. OEM versions of CP/M-86 were generally labelled Personal CP/M-86 after this, and they were somewhat different per manufacturer both in versioning and in feature set. As for the feature set, it would seem that each manufacturer sort of picked what they wanted from DRI’s software library and created a kind of distribution as we’d call it today.
The February 1983 issue of PC-Magazine ran an article “CP/M-86 Price Plunges to $60.” The article then speculated that this was likely due to ninety six percent of IBM’s PCs having shipped with PC-DOS and this most likely being due to CP/M’s price of $240 (or $753 in 2024) vs the PC-DOS price of $40 (or about $125 in 2024). The article then notes that CP/M-86 version 1.1 was the current version and that it shipped with a print spooler, graphics device drivers, and GSX (graphical system extensions). An important bit for us is that this article indicated that Digital Research was pivoting to direct to consumer sales and away from OEM. This is in-line with what I found elsewhere that DRI quickly ended development of CP/M-86 Plus to focus on CP/M-86 and Concurrent CP/M-86. This could be why features previously kept for the Plus versions were suddenly present in version 1.1. The first machine to get version 1.1 was the IBM PC, and beyond the extensions mentioned previously, this version supported hard disk drives.
MP/M-86 became Concurrent CP/M-86 in late 1982 with the merger of features from CP/M-86 1.1. Because DRI seems to have felt that versioning and naming were irrelevant, Concurrent CP/M-86 version 3 was the first version of CCP/M-86. Version 3.1 shipped in February of 1984. This version had a real time kernel and supported both single user and multiuser environments, networking (with Soft/Net providing LAN functions), shared libraries, GSX, IPC, and 8087 coprocessors. Reflecting the market at this point in time, version 3.1 also had a compatibility option that would allow a user to run most PC-DOS 1.1 (and 2.0 to some extent) applications, access PC formatted disks, and would allow up to four DOS applications to be multitasked. PC-MODE for Concurrent CP/M-86 became Concurrent DOS in May of 1984 which was essentially CCP/M-86 with PC-MODE integrated as a single software release according to DRI’s MicroNotes from May of 1984. It also dramatically expanded software compatibility with PC-DOS. In July of 1984, Concurrent DOS had become Concurrent PC DOS 3.2. This release fully integrated the DOS compatibility system into CP/M’s BIOS, BDOS, and CCP.
Concurrent DOS 68K made its debut in 1985, and this version was written in C. The last version was 1.21 in 1986. Concurrent DOS 286 was released in 1985, and it was also written in C. It’s final version was 1.2 in 1986. Both of these releases became FlexOS in 1986.
The Concurrent DOS (non-286 and non-68k) line up became Concurrent DOS XM and Concurrent DOS 386 starting in late 1986. Concurrent DOS XM allowed for bank switched memory of up to 8 MB in real mode. Concurrent DOS 386 was released in 1987 and obviously required a Intel 386. This version would run CP/M and DOS applications, supported multiple concurrent users, multitasking, and networking.
One interesting bit that I’ve struggled to find much information about was a UNIX port that DRI announced: UNIX System V for 286 with an applications library and CCP/M-86 compatibility. This was an official port by DRI on behalf of AT&T, Intel, and even Motorola. The Motorola involvement was a Concurrent DOS port to the M68000 along with all of DRI’s development tools, and this UNIX System V version. If anyone has more information about this UNIX port, please let me know.