Subject: Re: UNIX manpage history
From: Jerry Saltzer
Date: 06/11/2011 04:09
To: Kristaps Dzonsons

Kristaps,

Thanks for the progress report on the manpage history.  You have tracked
down quite a few pertinent facts and performed a valuable service in
straightening out the information and clearly outlining several
remaining mysteries.

In the interest of accuracy and completeness, I list below a few minor
details that you might consider adjusting.

					       Jerry

--------------- 1.  Memo, Modify and Ditto:  The documentation of those
commands (Computation Center Memo CC-205, March 11, 1963) was authored
by M. J. Leslie Lowry, Fernando J. Corbató, and J. Richard Steinberg.  I
believe it is appropriate to identify all three of them as the
designers.  A copy of CC-205 is attached below.

2.  The zip file reached by the link "source file containing RUNOFF"
also contains several other unrelated CTSS commands.  Attached below are
two files that contain just the source code (in MAD and FAP) of RUNOFF
and a text file showing how they relate.

3.  The file under the link "manual" leads to just a synopsis, not the
full manual.  Also, it appears to be a synopsis of the somewhat expanded
Multics version.

4.  The second source under 1964 is incorrectly attributed.  Patricia
Crisman was the general editor of the programmers guide.  I was the
author of section AH.9.01.  A better way to cite this source would be

• Jerome H. Saltzer, Manuscript Typing and Editing, section AH.9.01 of
The Compatible Time-Sharing System, A Programmers Guide, Patricia
Crisman, editor (second edition). M.I.T. Press, 1965 (revised December
1966). (HTML)

5.  The title of the 3rd source under 1964 has a typo; it should be
"TYPSET and RUNOFF,..."

6.  I just remembered another branch:  Starting sometime around 1966,
General Electric began running a commercial time-sharing service on the
GE-635, using Dartmouth's 635 version of DTSS.  I was surprised to
discover that RUNOFF was one of the applications available on that
service.  Whether that version of RUNOFF is somehow related to the
unconfirmed Bell Labs ports to the 635 is a mystery that needs some
followup.

7.  In the list of sources for the discussion of SCRIPT (1967) the link
"UNIX manpage history..." produces a 404 not found response.

8.  Under 1969 runoff (Douglas McIlroy) there is a comment that "the
mention of CTSS is incorrect."  There are two mentions of CTSS in the
previous few sentences; it is not clear which one you are identifying as
incorrect--I think it is the one in the context "under CTSS".  I suggest
that you clarify that.

However, I believe that your correction changing CTSS to GECOS is an
oversimplification.  There was an as yet unmentioned additional branch
in the RUNOFF tree:  an ASCII version of RUNOFF, named roff and written
in BCPL, running on CTSS.  Where this version fits in the tree I am not
sure.  It seems likely that there was just one roff BCPL source that was
recompiled everywhere that BCPL was available (GECOS, CTSS, and Multics
for starters) with support from local machine language interludes to the
underlying operating system.  So a better correction might be that the
BCPL roff was developed by McIlroy under GECOS and ported to CTSS and
Multics.

9.  Under 1971 roff (Ritchie) the 4th source is identified as "Multics
System Programmer's Manual".  It would be more accurate to identify it
as "Multics System Programmers Manual Table of Contents", so that people
don't expect to find the whole manual at the other end of that link.  A
few of the sections are there, but most are not.

10.  Under "people" the entry for my name is misleading--I was not
working in the CTSS project.  I wrote RUNOFF as a graduate student to
format my doctoral thesis proposal.