home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: vmsnet.pdp-11
- Path: sparky!uunet!utcsri!skule.ecf!torn!watserv2.uwaterloo.ca!watdragon.uwaterloo.ca!kcwellsc
- From: kcwellsc@watdragon.uwaterloo.ca (Ken Wellsch)
- Subject: Re: UNIX Operating System
- Message-ID: <Bxruxt.In2@watdragon.uwaterloo.ca>
- Organization: University of Waterloo
- References: <9211140841.AA28462@WLV.IIPO.GTEGSC.COM>
- Date: Sun, 15 Nov 1992 19:17:53 GMT
- Lines: 58
-
- According to "The UNIX Time-Sharing System" paper (I'm looking at a copy
- from "UNIX Programmer's Manual - Vol 2" from Bell Labs), there have been
- (as of V7 UNIX, 1978) four versions of UNIX:
-
- (1) circa 1969/70 running on PDP-7 and PDP-9's
- (2) circa 1971/72 running on a PDP-11/20
- (Dennis Richie creates C, the O/S is rewritten 1973)
- (3) circa 1974/75 running on PDP-11/40,45 (called V6)
- (4) circa 1978 running on PDP-11/70 (called V7)
-
- Version 6 UNIX was in fact distributed according to it's documentation
- on three RK05's or 9-track tape. Versions like "VENIX" I claim are
- actually derived from V6 (I use the filesystem as the major clue) and
- would thus run on the smaller address machines like the 11/40 (max 64Kb
- process size).
-
- Many people have seen true V7 shoe-horned onto a small address machine
- (e.g. 11/23) but as Steven said it is not a pretty sight. I recall
- a process limit possibly in the single digit range (for the whole system).
-
- UNIX was a research O/S (and I feel still is 8-) and the designers
- adapted to new hardware as it came out. Shifting to the larger
- process space (split I&D) machines was not at all surprising (take
- a look at how full your disks are and the old full+10% always seems
- to come true regardless of how much new mem/disk you buy).
-
- My feeling is the Berkeley folks early on when doing the 2bsd version
- wanted to counter the split-I&D aspect and developed the large instruction
- space model that allowed an enhanced V7 kernel to fit on a nonsplit
- machine.
-
- Of course AT&T licensing applies to all this software and as it costs
- lots of money (I recall $150,000 for a SysV source commercial license)
- the people who could afford it (commercial interests) went for the
- current and future hardware where the money can be made.
- (it would take such a commercial interest to also afford a binary
- resale license in order to even sell a copy they would have to
- develop for the PDP-11 - the cost isn't worth it to them I suspect)
-
- There is no denying that with a lot of effort, much software can be reduced
- dramatically in size requirements, but the people who can "legally" play
- with source don't need to fit things as small as 64Kb or 128Kb any more
- and the hobby people can't legally afford the licensing...
-
- It is almost ironic that of the three points the paper (referenced earlier)
- mention as "design influences", one of them is:
-
- "... there have always been fairly severe size constraints on the
- system and its software. Given the partially antagonistic desires
- for reasonable efficiency and expressive power, the size constraint
- has encouraged not only economy, but also a certain elegance of
- design. This may be a thinly disguised version of the ``salvation
- through suffering'' philosophy, but in our case it worked."
-
- But with all honesty, at some point it becomes practically impossible
- to have as severe a size constraint as a PDP-11 now represents in
- light of current hardware technology (obviously depending on application,
- there is still a lot of PDP-11 use today using DEC OS's)
-