The history of the IBM PC, part one: The deal of the century

Part 1a.

By this point Microsoft had also supplied the BASIC for Perkin-Elmer's very expensive Model 3600 "intelligent terminal".

These were more than just terminals capable of limited client-side processing for time-sharing systems. As stand alone desktop units they were powerful enough for scientists to acquire data, store it, process it and/or control data-intensive instruments which a year earlier had required a large non-desktop host, often a forklift model.

The 3600 was the first desktop to actually resemble an early IBM PC, horizontal unit having two prominent 5 1/2" floppies, and with the detachable keyboard (also introduced softkey F-row) & monitor.

Other than that the box was simply a microprocessor & memory with two RS-232 ports (known as serial COM ports ever since later release on IBM PC's), and an external instrument connection, not supposed to be expandable or upgradable internally.

ROM bios booted (without a floppy) to a novel Disk-Based Operating System known as the Perkin-Elmer Terminal Operating System (PETOS) where a few of the DOS commands were available from ROM, but the remaining majority of the commands were expected to be present on a floppy residing at DISK0. DISK1 was expected to usually be employed for application program & data storage.

These were proprietary Perkin-Elmer programs to interface with their own scientific instruments but many users wanted to develop their own programs in BASIC like you could on the competitive Hewlett-Packard equipment using their built-in HP BASIC.

The team that had designed the 3600 had of course been separated into numerous more rewarding projects.

Anyway Perkin-Elmer got Microsoft to provide the BASIC for the 3600, and and as we all gained deeper knowledge of PETOS by operating this equipment, it really helped later when the IBM PC was launched because its DOS had such an uncanny similarity.

"Kildall, meanwhile, often didn't even seem certain he wanted to be running a business in the first place:"

I should mention Gates wrote the Open Letter to Hobbyists. Of course, this was before "open source" or "free software" even was a term. At the time, CP/M had a "BDOS" supplied by DRI with the OEM having to write the "BIOS". There were firms such like Lifeboat which had writing the BIOS as part of their job.

Thinking about it, I wonder if it would have been feasible for Kildall to work on CP/M full-time at Intel and release CP/M source to public domain instead. It would be nice if CP/M was the first thing ported to 8086 back in 1978 for example.

The IBM PC used the 8088 CPU which offered the same 16-bit execution unit of the 8086 (and thus superior to the 8-bit Apple II) while using only an 8-bit bus which was the same size bus as the Apple II. The 8-bit bus of the 8088 was compatible with the more mature 8085 thus allowing the IBM PC to use more proven, cheaper parts. The 8088 part itself was cheaper than the 8086.

In my view, this was a brilliant design choice and a pity that Apple didn't do the same.

The 8088 was designed in the Haifa Israeli Design Center.

There are a very large number of key points in this story which directly contradict previous published accounts - and no bibliography. Color me flummoxed.

"It was a self-contained, Turing complete, programmable machine no larger than a suitcase" - False

Languages can be Turing complete. ISAs (Instruction Set Architectures) can be Turing complete. Machines... um, no, not of present design. That infinite tape provides infinite memory including infinite register space. Hard to duplicate with a real machine. There's a limit to the largest number you could square on an IBM PC, so it's not Turing complete - in other words, there are functions it can't calculate, even given infinite time.