[pups] The Software Tools Virtual Operating System

Bill Gunshannon bill at cs.scranton.edu
Fri Nov 2 11:53:13 AEST 2001

On Thu, 1 Nov 2001, David C. Jenner wrote:

> Bill,
> I can assure you that what you were pointed to at the Ken Yap's Links
> is what you are seeking.  It may not be the final version, but it is
> essentially what I received on a tape 20 years ago.  There's even some
> later material circa 1983, which postdated what I had, in the "Toys" file.

Hmmmm.  I'm beginning to think as it neared the end and interest trailed
off much of the work did not get rolled back in to the base distribution.
My last experience was around 84-85 and it was a package from GA Tech
that ranon the Prime 50 series minis.  It was alot like using Eunice on
a VAX/VMS system.  A quick scan of the stuff from Ken Yap found no mention
of a number of systems that were known to exist by that time.  No PDP-11,
no Prime, some mention of the VAX, no mention of Unix (don't ask me why,
but the VOS was ported to Unix!!).  I guess what I need to do is see what
systems are supported in what I have and try to get a system up again to
check out.

> I probably still have the tape, but, it's 20 years old, hasn't been
> used for at least 15 years, who knows what condition it's in, and it's
> not much different than the content on the Web.

What are the odds that a 15 year old tape is even readable today??  I know
when I found the original BSD tapes here even with their being stored in the
computer room, they were unreadable.

> You must realize, I guess, that you aren't getting a complete operating
> system when you speak of STVOS, but just a ratfor translator and sources
> for lots of Unix-like utilities.

And primitives to translate between the host OS and the VOS API.

>                                  You need to supply a Fortran compiler
> and operating system on which to build this.

And many were supported.  I was hoping to find enough of them to have a
good example of what problems were run into doing the ports.  Can go a
long way in helping with other porting efforts.  And because the API is
very Unix-like it offers some intersting possibilities for expansion.

>                                               The idea was to make a
> highly-portable set of software development tools and utilities that
> could be ported and used across many OSes, thus making what you develop
> available across many OSes.  There is no OS (i.e., resource management,
> file system, etc.) included.

True.  But a common API with hooks into a number of very different OSes.

> What you refer to in your emails about VOS bears little resemblance
> to the STVOS, because STVOS wasn't an OS.

Probably depends on your definition of OS.  It was an ambitious project
at the time and an idea whose time may just now be coming into vogue.

>                                            Today's VOSes, like a Java
> machine, are at least one step beyond the STVOS.

Again, I am not sure I agree.  To me the Java VM is just the UCSD P-machine
warmed over.  One of the reasons things like the P-Machine and VOSes didn't
fly 20 years ago was performance.  We were trying to wring every last bit
(no pun intended!) of performance out of our hardware.  We frequently still
did a lot of our programming in assembler (I was doing things like Prime
50-series and Univac-1100 assembler and almost anything on a micro was either
complete or heavily laced with assembler, LSI-11, Z80, M68K.)  Today, for all
intents and purposes we have cpu cycles to burn.  Look at the popularity of
hardware emulators.  E11, Charon VAX, SIMH.  And people talking about how
these emulators outperform the real hardware and could easily be used as
production systems.  Maybe it's time to look into reviving some of these
ideas, but hopefully, not with a return to the beginning and a total re-
invention of the wheel.

> You might want to establish what the final date of release of STVOS
> was to determine what the final version was.  As I recall, it wasn't
> too much later than 1981.  (The Toys tape is 1983.)  Somewhere I have
> a pile of old newsletters that would have the answer, but they're
> boxed away in storage.  I won't be able to dig for them for another
> month.

I know the Prime version was still available until the mid 80's. but much
of this may have been independant work as STUG may have already faded
into the background.

I guess the thing that bothers me the most is not wether or not this can be
turned into something usable, but the fact that what was an impressive work
for the time it was done has been allowed to all but disappear.  Maybe I'm
getting maudlin in my old age.  :-)

I appreciate everyone's help and as I said previously, it makes me appreciate
even more the work of Warren and PUPS and Tim Shoppa as well.

All the best.


Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolves
bill at cs.scranton.edu     |  and a sheep voting on what's for dinner.
University of Scranton   |
Scranton, Pennsylvania   |         #include <std.disclaimer.h>

More information about the TUHS mailing list