Subject: Re: Unix manpage history: contact for Bill Joy and Cynthia Livingston?
From: Kirk McKusick
Date: 20/03/2026, 23:15
To: Kristaps Dzonsons

> > From: Kristaps Dzonsons
> > To: Kirk McKusick
> > Subject: Re: Unix manpage history: contact for Bill Joy and Cynthia Livingston?
> > Date: Fri, 20 Mar 2026 21:42:32 -0700
> > 
> > Kirk,
> > 
>> >> Keith lead the effort to get redistributable versions of the AT&T
>> >> libraries and utilities. He would get up at the Usenix Birds-of-a-Feather
>> >> session at which we were talking about the latest things happening
>> >> in BSD and put up a list of the utilities and library functions
>> >> that we needed to have rewritten. He would ask folks to contribute
>> >> to the effort by writing new implementations from scratch that
>> >> could be freely redistributed. Most of the contributions required
>> >> that he seriously rewrite them to make them usable. I remember asking
>> >> him why he didn't add his name as one of the authors and his reply
>> >> was that if AT&T was unhappy about the rewritten program he wanted
>> >> them to know to whom to complain.
> > 
> > Would you mind if I include these messages (redacted for privacy, of 
> > coures) in the record?  This is exactly the kind of information I need!

I am fine with your including these messages.

>> >> Looking at the SCCS log however, it appears that Keith was the original
>> >> author of the man command:
>> >> 
>> >> D 5.1 87/06/29 20:53:50 bostic 1 0
>> >> date and time created 87/06/29 20:53:50 by bostic
>> >> 
>> >> So, as a Berkeley written command, it could be freely redistributed.

> > In the 1BSD sources, 
> > https://github.com/dspinellis/unix-history-repo/blob/BSD-1/s6/man.c, I see:
> > 
> >   * man - intelligent man command
> >   * Author: Bill Joy UCB August 25, 1977

I looked in my 1bsd archive and everything is archived in cont.a
files which file reports as `old 16-bit-int little-endian archive'.
I have no idea how to extract their contents, but I assume that
the archive at which you are looking has done that.

> > This looks like the first C version of man(1), as Ken Thompson's was 
> > just a shell script.  The manwhere utility is also in 1BSD, 
> > https://github.com/dspinellis/unix-history-repo/blob/BSD-1/s6/manwhere.c:
> > 
> >   * manwhere - where is something in the manual
> >   * Author: Bill Joy UCB August 29, 1977
> > 
> > Then in 2BSD, manwhere is removed and whatis arrives:
> > 
> >   * whatis - what the heck is that file anyways
> >   * Bill Joy UCB
> > 
> > apropos.c is also there with the same 1979 copyright, but no message as 
> > to who wrote it.  Same with makewhatis.sh.  I presume those were Bill as 
> > well, but I don't have the log, just snapshots.

I concur with your guesses above. We had no source control until
SCCS showed up in 1980 and the project began using it (three years
after Bill Joy began his initial Berkeley UNIX development). 
Initially only the kernel was put under SCCS, but by 1983 everything
had been put under SCCS.

> > In the CSRG repo 
> > (https://github.com/csrg/csrg/commits/master/usr.bin/whatis/whatis.c), I 
> > presume Bostic's 1987 version is the re-write to shed AT&T bits?  And 
> > same with the 1988 apropos/whatis split?

I concur with your analysis.

> > Cynthia has responded already, by the way!  I'm really happy to get her 
> > on the record, as manpages wouldn't be where they are today without mdoc.

Yes, her work was a big step foprward.

> > That reminds me: you looked to have imported the -man macros from 4.1BSD 
> > into old/tmac/tmac.an.  Do you know who (or if anybody?) re-wrote Doug 
> > McIlroy's original AT&T macros under the BSD license?

I believe that the plan at least was that all the manual pages that
were to be freely redistributed were written using Cynthia's mdoc
macros. I don't believe that McIlroy's original AT&T macros were ever
rewritten under the BSD license.