Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: version B 2.10.2 9/18/84; site brl-tgr.ARPA Path: utzoo!watmath!clyde!cbosgd!ihnp4!mhuxn!mhuxm!sftig!sftri!sfmag!eagle!ulysses!allegra!bellcore!decvax!wanginst!ucadmus!harvard!seismo!brl-tgr!tgr!Munck@MITRE-BEDFORD.ARPA From: Munck@MITRE-BEDFORD.ARPA Newsgroups: net.unix-wizards Subject: VM/370 origins Message-ID: <6912@brl-tgr.ARPA> Date: Wed, 2-Jan-85 14:56:58 EST Article-I.D.: brl-tgr.6912 Posted: Wed Jan 2 14:56:58 1985 Date-Received: Fri, 4-Jan-85 04:53:14 EST Sender: news@brl-tgr.ARPA Organization: Ballistic Research Lab Lines: 51 David Green and Andrew Klossner have set most of the facts straight, but there is still some history missing or wrong: VM/370 originated as CP/40, a small internal project at the IBM Cambridge Scientific Center, 540 Tech Square, Cambridge, MA in about 1964. In that same building, MIT people were constructing MULTICS, of which UNIX is an inferior descendent; hence this use of UNIX-WIZARDS. CP/40 was intended to provide resources for the research in operating systems being done at the Scientific Center. It used a /360 mod 40 with a patched-on hardware relocation box to provide virtual bare /360s to each user. (The idea was an outgrowth of earlier work done by R.W O'Neill on the 70XX machine at Watson Research Center in the late 50s.) To aid programming, a simple interactive executive was built supporting little more than an assembler, linker, editor, and file system: CMS. That first system was pretty feeble; I can remember groans sounding up and down the corridor whenever anyone did an assembly. When the /360 mod 67 came out with hardware relocation (virtual memory) standard, the Cambridge people immediately moved to the more powerful machine, creating CP-67/CMS. IBM's flagship OS for that machine, TSS, was late and buggy, causing a number of installations to acquire CP for their 67. (The Univ. of Michigan built their own very nice system, MTS.) We started running CP at Brown Univ. in 1969 after a great debate over whether interactive computing was a fad. About this same time, some of the original developers of CP (Dick Bayless, Mike Field, others) left IBM to start their own company, CSS (later National CSS, later NCSS, finally bought out by Dun and Bradstreet). It was one of the leaders in the heyday of phone-in interactive mainframe computing. At Brown, we taught hands-on operating system design and graphics on CP to several generations of students who have gone on to prominence in the field. (U. of Mich. produced a similar vintage group.) By the time the /370 line was announced, there were about 50 CP-67 installations, mostly universities. (Less than 100 67s were built.) We were amused by IBM's "invention" of virtual memory on the /370, as were Burroughs and GE. CP/CMS made the jump to the /370 line easily, becoming VM/370, and also began to rival IBM's "main" operating system, OS/360-MVT -> MVS. I've heard rumors that pitched political battles between the two factions at high levels in IBM continue to the present day. Though I have no figures, I'd bet that there are more VM users than UNIX users in the world today. Finally, CP is now available on my desktop at a TOTAL price of $5500 (with MITRE's giant discount) in a PC/AT with the /370 board. I can't wait to see if any of my code has survived the decades. -- Bob Munck