PWB/UNIX
rdave...@gtewis.arpa
I've recently acquired a copy of the PWB/UNIX User's Manual in two volumes,
(one section looks like the current User's Manual and the other like the
Programmer's Manual) printed in May 1977. Can anyone tell me where PWB/UNIX
fits in to the UNIX history? I've never heard of it before.
thanks,
rob
_-----------------------------------------------------------------_
| Robertson G. Davenport |
| GTE Govt. Systems Corp. |
| Billerica, MA C the world.... |
| BELL : (617) 671-5180 |
| ARPA : rdave...@gtewis.arpa |
| CIS : 73407,3716 |
| UUCP : ...!{harvard,rutgers,uunet,ulowell}!gtewis.arpa!rdavenport |
\___________________________________________________________________/
"If you want to eat hippopautamus, you've got to pay the freight.";
-- some IBM guy
-----
Ron Natalie
Workbench) saw the light of day (out side of ATT) around 1978 and was
developed by the group that eventually evolved into the people who
were responsible for PWBII, System 3, and System V. It's main raison
d'etre was the enhanced ability to keep track of programs to be down
loaded to other computer environments (like IBM RJE). SCCS has it's
roots here. Originally it was a slightly bug fixed V6 kernel with
a couple of enhancements (like the empty() system call) and featured
a few extra niceties like the PWB shell, which was distinguished from
the existing shell by having things like shell variables and a user
setable path.
-Ron
Alan Matsuoka
>
> Can anyone tell me where PWB/UNIX
> fits in to the UNIX history? I've never heard of it before.
>
[ somebody please correct me if I am wrong anywhere... This goes
back inmy memory a ways].
I worked on a PWB system about 9 years ago. As I can recall, it was a
sort of Version 6.5 ( At least the C compiler was ). Version 6.5 was
released to a number of universities ( I think... I came from a Version 6
site [ or 6th Edition as it was called ]). PWB was a sort of beefed
up version that came with SCCS.
It had the 6th Edition I node format ( with some modifications [I forget] ).
We didn't have such nice things like fsck to fix things up when things
crashed, just fsdb , a 'file system debugger' that was considered a
great help considering that all we had with V6 was rm, clri, dcheck, icheck
and ncheck to fix up the file system. A crash took a while to recover from.
There were the old restrictions on file sizes like in V6.
There were restrictions in the C language at the time,
things like no enums, voids, arrays of unions.
unions, typedefs, and unsigned math were new back then.
(How many of you can remember that unsigned arithmetic was done
using char * in V6?).
Last but not least, I believe that it ran only on PDP11/70's with Floating
Point. There was some bizarre reason why you needed the floating point unit
but you could live without it if you had a source license (everybody did
back then).
Oh, yes.. Then there was the Mashey shell. This was a shell that appeared
only with PWB.
Now most people, (mainly institutions) waited until Version 7 came
out since it was a lot better. After V7, Western Electric changed the
naming scheme and came out with System III. A sort of PWD, V7 and other
stuff put together....
--------------------------------------------------------------------------------
Alan Matsuoka, Develcon Electronics, 515 Consumers Road,
Suite 500 Willowdale, ON Canada M2J 4Z2
uucp: {decvax,ihnp4,utai,uunet}!watmath!dvlmarv!alanm phone: (416) 495-8666
{utai,mnetor}!lsuc!dvlmarv!alanm
sask!zaphod!dvlmarv!alanm
Ron Natalie
even require split-I/D so you could run it on 11/34's and such provided they
had memory management.
What I believe you are referring to as 6.5 is the phototypesetter C compiler.
When you bought troff, you got a fixed up, pre-version 7 compiler that had
things like the equals-op operators the right way around. There was also a
kernel diff listing circulating around referred to as "fifty fixes."
There was actually a PWB II release which had a V7 kernel as a base, if I
recall.
-Ron
Andrew Tannenbaum
> (one section looks like the current User's Manual and the other like the
> Programmer's Manual) printed in May 1977. Can anyone tell me where PWB/UNIX
> fits in to the UNIX history? I've never heard of it before.
There were two major versions of PWB/UNIX, 1.0 was V6 based, 2.0 was v7
based. The PWB (Programmer's Workbench) group were UNIX developers
whose major charter was to develop economical UNIX systems as front
ends to real computers (like IBM, Univac, Honeywell), so that a system
analyst could have one terminal on his desk instead of three terminals
(or cardreaders), he could use one editor - one integrated set of tools.
Some of the major contributions of PWB are the MM macros for troff and
the UNIX RJE stuff (which was heavily used in Bell Labs, at least when
I was there - until 1983). See the paper "The Programmer's Workbench"
by Dolotta, Haight, and Mashey, in the classic UNIX BSTJ from the
summer of 1978.
While at MASSCOMP, I drew up this timeline of UNIX history
until AT&T SVR2/4.2BSD.
Andrew Tannenbaum Interactive Boston, MA +1 617 247 1155
<<>><<>><<>><<>><<>><<>><<>><<>><<>><<>><<>><<>><<>><<>><<>>
'\" tbl | troff
'\" need a blank line to init troff state
.sp .5i
.TS
box;
r l lbp+2
r l||l.
Great Moments in UNIX System History
=
1969 Thompson fiddles with filesystem ideas on Multics
Thompson ports "space travel" from MULTICS to
GECOS to PDP-7, hacks UNIX filesystem and...
UNIX V1 written in assembly by Thompson for PDP-7&9
_
1970 Thompson intends to write FORTRAN using TMG, writes B instead.
After being denied for $.5M for a research DEC-10, Ossanna
asks for $65K for a new PDP-11/20 for text processing research.
Ordered in May, comes in August
Runs UNIX in core until disk comes in December
Kernighan suggests name UNIX
_
1971 Feb UNIX V2 PDP-11/20 in assembly language on unprotected hardware
Work begins on C
Summer BTL patent dept runs text processing on 11/20 UNIX
Nov UNIX V1 manual
_
1972 Jun UNIX V2 manual
UNIX gets pipes
_
1973 Feb V3 Manual
Feb UNIX V3 PDP-11/34 40 45 60 70 multiprogramming
C runs on Honeywell 6000, IBM 370
Oct UNIX first mentioned in public at ACM 4th SOSP
Snyder Portable C MIT Masters Thesis
PWB/UNIX conceived - RJE SCCS PWB/MM
MERT designed for PDP-11/45
_
1974 June V4 Manual
July CACM UNIX paper "The UNIX Timesharing System"
UNIX V5 university release
_
1975 May UNIX V6 manual
_
1976 UNIX V6 commercial release
LSX LSI-11 UNIX (Minimum Configuration ~$7000)
Feasibility study for UNIX under VM/370 at Princeton
_
1977 early Work begins on Interdata 8/32 UNIX port at Bell Labs
Interdata 7/32 UNIX port at at U Wollongong Australia
UNIX V6 commercial sublicenses
USG UNIX Generic Issue 3 (UNIX V6 based)
May PWB/UNIX 1.0 (UNIX V6 based) Bell Release
Jun John Lions' commentary on UNIX V6
late MERT Release 0
_
1978 Spring Johnson and Ritchie Interdata (later PE) port complete
Jul Bell System Tech Journal UNIX issue, Part 2 Vol. 57 No. 6
current: USG UNIX, PWB/UNIX, UNIX V6, Research UNIX
UNIX TS 1.0 (UNIX V7 based) Bell release
UNIX 32V VAX port (UNIX V7 based)
_
1979 Jan UNIX V7 manual
Apr UNIX RT (supersedes MERT)
Jun PWB/UNIX 2.0 (UNIX TS based) Bell release
Dec 3BSD UNIX VAX virtual memory
_
1980 Feb UNIX TS 1.3 VAX & PDP-11 Bell release
Apr UNIVAC 1100 UNIX TS 2.0+ Bell release
Jun UNIX TS 2.0 VAX only Bell release
Jun UNIX 3.0 (replaces TS and PWB) Bell release
Oct 4.0BSD UNIX
_
1981 March UNIX/370 Bell Release
Mar 3B UNIX 3.0 Bell Release
Jun 4.1BSD UNIX
_
1982 Jun UNIX 5.0 Bell release
_
1983 Aug 4.2BSD manual
Dec UNIX System V Release 2 manual
_
.T&
r r r.
Andy Tannenbaum MASSCOMP 9/84
.TE
John F. Haugh II
summer of 1981. as far as i know it was only released internally and to
some universities which had professors on loan from the labs.
i had the misfortune of having to deal with it for a short time, prior
to xavier (where i hung out) getting release 5.0 running on the 11/45.
i can't say much about 4.0, except that it appeared to be closer to
5.0 than to 3.1 (did i get that right??? was it 3.2???) we were having
hardware trouble so no one got to use the system much.
as i recall, 4.0 still fit in 128K or so of memory. release 5.0 had to
be trimmed down quite a ways, and it still took more than 128K. which
leads to the next question. has anyone every managed to fit s5r2 on a
pdp-11?
- john.
--
John F. Haugh II +--------- Cute Chocolate Quote ---------
HASA, "S" Division | "USENET should not be confused with
UUCP: killer!rpp386!jfh | something that matters, like CHOCOLATE"
DOMAIN: j...@rpp386.uucp | -- with my apologizes
Ken Keirnan
>andy didn't include release 4.0, which as i recall was released in the
>summer of 1981. as far as i know it was only released internally and to
>some universities which had professors on loan from the labs.
>
>i had the misfortune of having to deal with it for a short time, prior
>to xavier (where i hung out) getting release 5.0 running on the 11/45.
>i can't say much about 4.0, except that it appeared to be closer to
>5.0 than to 3.1 (did i get that right??? was it 3.2???) we were having
>hardware trouble so no one got to use the system much.
>
>as i recall, 4.0 still fit in 128K or so of memory. release 5.0 had to
>be trimmed down quite a ways, and it still took more than 128K. which
>leads to the next question. has anyone every managed to fit s5r2 on a
>pdp-11?
First, you are correct about UNIX 4.0. This version was something of a
"pre-release" of System V that was missing several major enhancements that
eventually appeared in 5.0 (I don't remember the specifics), and yes, this
version was only released within the Bell System (I still have 4.0 running
on 2 pdp-11/70s).
Second, System V release 2 was the last release of UNIX from Bell Labs to
run on the pdp-11/70 (I don't know if it runs on the 11/45). The pdp-11
version of release 2 is missing several operating system features and
utilities that are available on larger machines, but the "Labs" managed
to fit the kernel text into 64K. Kernel data is larger than 64K, but is
handled by using a combination of pdp-11's "kernel" and "supervisory"
modes to map the necessary data space. In spite of all the memory
management manipulation required, Sys VR2 runs great on the pdp-11/70.
Yes, I still have an 11/70 running SVR2, and I still find it a challange
to shoehorn large programs on to the system (Documentors Work Bench 2.0
was the most recent).
Ken Keirnan
--
Ken Keirnan - Pacific Bell - {att,bellcore,sun,ames,pyramid}!pacbell!pbhyf!kjk
San Ramon, California k...@pbhyf.PacBell.COM