[TUHS] "A New Command Language for UNIX and Related Systems", Sturzenbecker CL (1982)

[TUHS] "A New Command Language for UNIX and Related Systems", Sturzenbecker CL (1982)

[Alex Bochannek via TUHS]

Hello!

Last week I got a scan of Bell Labs TM-82-45192-3, M. C. Sturzenbecker,
"A New Command Language for UNIX and Related Systems" from the Bell Labs
archives thanks to Rebecca Nadolny at Nokia. Rich Salz suggested I
should share the information here.

I first noticed a reference to this document in David Korn's
"Introduction to KSH-i" memo from 1986 and because I had not heard about
it before, requested a copy.

Most significantly, this Shell uses typed variables and goes into some
detail explaining the challenges around that. The references to broader
developments in command languages and the desire to make this CL
applicable to Unix and non-Unix systems is quite interesting.

I am curious how well known this memo was at the time given the broad
distribution list it was apparently sent to (which I suspect was common
for TMs). David Korn cites it, but were others aware of it as well? Was
it known outside Bell Labs? Was the implementation of CL used by others?
Was it ever implemented on non-Unix systems? The document mentions that
CL was implemented for Unix Release 4.0, which also made me wonder how
widely that release was used inside the Bell System.

Unfortunately the scan is too large to send here, but I am happy to
share it.

Thanks!

-- 
Alex.

[TUHS] Re: "A New Command Language for UNIX and Related Systems", Sturzenbecker CL (1982)

[Arnold Robbins via TUHS]

Hi.

Alex Bochannek via TUHS <tuhs@tuhs.org> wrote:

> The document mentions that
> CL was implemented for Unix Release 4.0, which also made me wonder how
> widely that release was used inside the Bell System.

I think it was used pretty widely. I've described this before, check
the archives, but circa 19833-1984 I did some contract programming
for Southern Bell on a PDP-11/70 running Unix 4.0.  It was the
standard version of Unix inside the Bell System.

Arnold

[TUHS] Re: "A New Command Language for UNIX and Related Systems", Sturzenbecker CL (1982)

[segaloco via TUHS]

On Tuesday, February 24th, 2026 at 12:47, Arnold Robbins via TUHS <tuhs@tuhs.org> wrote:

> Hi.
> 
> Alex Bochannek via TUHS <tuhs@tuhs.org> wrote:
> 
> > The document mentions that
> > CL was implemented for Unix Release 4.0, which also made me wonder how
> > widely that release was used inside the Bell System.
> 
> I think it was used pretty widely. I've described this before, check
> the archives, but circa 19833-1984 I did some contract programming
> for Southern Bell on a PDP-11/70 running Unix 4.0.  It was the
> standard version of Unix inside the Bell System.
> 
> Arnold
> 

Long musing on Release 4.x incoming.

My reading of documents agrees that Release 4.x was pretty wide-spread.
The story I've heard goes that AT&T tended to stabilize with odd
releases and then add a bunch of new stuff on even releases, so BTL was
probably more likely to be on a even release when available, same story
from what I've read with PWB 2.0.  I've read anecdotes that SVR2 was
originally to be Release 6.0 but by that time divested AT&T decided to
just go all in on UNIX releases and stuck with the System V trademark.
Internal BTL editions of this release are indeed also labeled System V
Release 2 whereas the prior versions were Release 5.0 for internal use.

Release 4.0 was popular enough to have a couple of nice printed volumes
of papers from the "Documents For" collection.  One such starter package
is available here:

https://archive.org/details/unix-programming-starter-package

It was also the beginning of the end for the simplicity of document
distro, Release 4.x is the first release we see the manual split into
u_man and a_man, and the unbundling of the secondary volume TROFF
sources may have occurred around the same time.  By SVR2, what was a
single programmer's manual and volume 2 is now a_man, u_man, and p_man,
along with all the papers strewn amongst a number of different
independent volumes.

However unlikely, Release 4.x is potentially more likely to have
survived in a cast-off tape or disk pack than earlier PWB and USG stuff.
However unlikely one of my goals is to hopefully start finding rapport
with nearby telco offices to see if I can suss out any old stuff from
that time.  That said, I've got no idea how wide-spread the various BTL
UNIX flavors were outside in the larger Bell System ecosystem, if for
instance any random telco office would've had USG generics or PWB, or if
most of their UNIX would be in more deeply embedded apps like COSMOS or
ESS units.  I don't know how this bodes for the chances of a UNIX 4.0 or
any other esoteric thing just sitting and rotting under the floor of a
switching office somewhere.  Unfortunately my letter and gift of an old
Bell System commemorative token have resulted in no return of contact
from the closest office to me (360-REgent) so remains a mystery.  They
do have what looks like a pre-divestiture 5ESS running in their
basement at least...so maybe there's hope for some old UNIX hiding
somewhere...

- Matt G.

[TUHS] Re: "A New Command Language for UNIX and Related Systems", Sturzenbecker CL (1982)

[Arnold Robbins via TUHS]

The whole point of USG -- the Unix Support Group -- was to unify the
various Unix variants in use in the Bell System into one flavor
that could be used everywhere.  Thus the fact that there are bits
and pieces from Research, PWB, and even "Columbus Unix" in Unix 3.0
and 4.0 and on to System V.

In my work at Southern Bell, Unix was used for generic computing,
not just software development. My job was to write a program that
would nightly read data read in from an HR tape and construct a
fast access database such that guards in the lobby could look people
up by name.

My point is that Unix wasn't just in the switching centers.

As to releases, odd vs. even, etc., I don't know about that. I was
told that AT&T released to the world one version behind what they
ran internally, thus Unix 4.0 inside the Bell System, Unix 3.0 (System III)
outside.  With divestiture on the way and AT&T getting ready
to go into the computer business, they decided to skip releasing
Unix 4.0 and go straight to System V both internally and externally.

HTH,

Arnold

segaloco via TUHS <tuhs@tuhs.org> wrote:

> Long musing on Release 4.x incoming.
>
> My reading of documents agrees that Release 4.x was pretty wide-spread.
> The story I've heard goes that AT&T tended to stabilize with odd
> releases and then add a bunch of new stuff on even releases, so BTL was
> probably more likely to be on a even release when available, same story
> from what I've read with PWB 2.0.  I've read anecdotes that SVR2 was
> originally to be Release 6.0 but by that time divested AT&T decided to
> just go all in on UNIX releases and stuck with the System V trademark.
> Internal BTL editions of this release are indeed also labeled System V
> Release 2 whereas the prior versions were Release 5.0 for internal use.
>
> Release 4.0 was popular enough to have a couple of nice printed volumes
> of papers from the "Documents For" collection.  One such starter package
> is available here:
>
> https://archive.org/details/unix-programming-starter-package
>
> It was also the beginning of the end for the simplicity of document
> distro, Release 4.x is the first release we see the manual split into
> u_man and a_man, and the unbundling of the secondary volume TROFF
> sources may have occurred around the same time.  By SVR2, what was a
> single programmer's manual and volume 2 is now a_man, u_man, and p_man,
> along with all the papers strewn amongst a number of different
> independent volumes.
>
> However unlikely, Release 4.x is potentially more likely to have
> survived in a cast-off tape or disk pack than earlier PWB and USG stuff.
> However unlikely one of my goals is to hopefully start finding rapport
> with nearby telco offices to see if I can suss out any old stuff from
> that time.  That said, I've got no idea how wide-spread the various BTL
> UNIX flavors were outside in the larger Bell System ecosystem, if for
> instance any random telco office would've had USG generics or PWB, or if
> most of their UNIX would be in more deeply embedded apps like COSMOS or
> ESS units.  I don't know how this bodes for the chances of a UNIX 4.0 or
> any other esoteric thing just sitting and rotting under the floor of a
> switching office somewhere.  Unfortunately my letter and gift of an old
> Bell System commemorative token have resulted in no return of contact
> from the closest office to me (360-REgent) so remains a mystery.  They
> do have what looks like a pre-divestiture 5ESS running in their
> basement at least...so maybe there's hope for some old UNIX hiding
> somewhere...
>
> - Matt G.