partial installation of DNA/EMU Fortran

[David Kennedy]

cross posted on Discord
I was gifted a thumb drive containing a scan of the DNA Systems Fortran IV manual.

Besides the compiler and subroutines, the DNA doc mentions that there were accompanying decks to replace the IBM CLB, DUP, and Supervisor phase 6 as well.

In high school, our 1130 ran the EMU Fortran with its subroutine library, but as far as I can recall, the other three decks were never a part of the build.
Does anyone know what functionality is lost if only the compiler and subroutines are used?

From my cursory reading, there appears to be considerations around the Floating point ACcumulator and perhaps some Fortran control card processing.

David

[Carlos Abramo]

David, 

Do you mind sharing the link to it with us? 

Thanks a lot 

Carlos 

--
IBM1130 discord channel: https://discord.gg/nrFEMt48mE
---
You received this message because you are subscribed to the Google Groups "IBM1130" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ibm1130+u...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/ibm1130/69b562da-4786-4da9-ab9c-0678b6b0817an%40googlegroups.com.

[David Kennedy]

It's not on any website, just the thumb drive.
Where can I upload it?  I don't have much online presence, just email.

David

[Peter Diehr]

Hello, David!

The Fortran IV compiler under discussion was originally written by
myself, along with C. David Morse, both working at Eastern Michigan
University, at the Instructional Computer Systems group. We started in
the summer of 1971, had a working system by October, and our first
distributions in November. Additional work was done during the winter
semester, concluding by late March of 1972. We worked a lot of long
hours, including nights and weekends - so that we could get the
required computer time. I was completing my senior year, while David
was a graduate student in the Math department. I had been working at
ICS since the summer of 1969, shortly after my discharge from the
service, and was the senior programmer. David had started in the fall
of 1970, IIRC.

I was the primary author of the EMU-IBM Fortran IV manual, which was
written in two phases: October 1971 for the "initial version", and
March 1972 for the "final" version, which was widely distributed.
David reviewed everything I wrote, and developed the technical
appendices that describe the load decks, etc. We distributed the
EMU-IBM Fortran IV compiler for essentially distribution costs, but
announced that we would no longer be supporting it, as the IBM 1130
computers were being phased out that summer. The product and source
code were made public domain. I had written all of the parsing and
syntax checking, and the generation of the symbol table; David handled
the code generation, and provided the new required functions, while I
wrapped up with the printing, including the interleaved error
messages, and also wrote all of the error messages. I also wrote most
of the debugging tools.

In April of 1972, David went to work for DNA Systems. We had been
collaborating with them on some features for the past few months. I
completed my share of that work, and went on to graduate school, for
my MBA. DNA Systems then began selling the Fortran IV compiler, and
may have made some further changes. They would have added some of
their useful IBM 1130 utilities to the package, which explains the
presence of DUP in the installation. I don't recall what "CLB" stands
for; SUP, phase 6 is probably DNA Systems 1130 DOS changes; we also
had a bunch at ICS, which I had written, but though they were
convenient for us, they were not required for the Fortran compiler.

I would like to compare the final version of the EMU-IBM Fortran IV
manual with the DNA Systems version. I recently donated my copy to
the EMU Library, archives section, as part of the history of ICS,
along with documentation and photos I've had for the past fifty-plus
years! If you can send me your file(s), I can read through and
compare (by eye), and summarize the differences. Thanks in advance!


Best regards, Peter Diehr, PhD, MBA

[David Kennedy]

A little bit of RTFM and
I think I have the permissions right.
https://drive.google.com/file/d/1SPinoLQy3yZkRvNalGvhq44dZtagWlFm/view?usp=drive_link
David

On Monday, August 10, 2026 at 11:10:42 AM UTC-4 carlos...@gmail.com wrote:

[Peter Diehr]

Google drive link works.

Best regards, Peter

> To view this discussion visit https://groups.google.com/d/msgid/ibm1130/70779a81-674a-4a69-8a46-05ee4d6f8891n%40googlegroups.com.

[David Kennedy]

After spending some time reading the manuals on bitsavers.org, I'm slowly recalling memories from 50 years ago.
As far as I can remember, the EMU compiler and subroutine libraries came with very little documentation, and so I don't think anyone made use of any features documented in the DNA manual other than logical variables, expressions, and the logical IF statement.
The DNA manual references complex variables, named COMMON, two word integers with values up to 2**31, REREAD, and more.
Named COMMON and complex variables probably explain the updated Core Load Builder (CLB) and Supervisor phase 6 (which I think handles the core dump requests.
The DNA manual also refers to the V2M13 DUP which I assume is from IBM and not DNA.
If programs had no complex variables and no named common, perhaps the IBM CLB could build the core image?
Further reading required.
David

[Carl Claunch]

DUP is a standard part of the Disk Monitor System V2 on all 1130 systems. 

[Ron Carlson]

Works with Google screen reader, printing seems to be disabled.

Any way it can be printed for reading, making notes?

Thanks, Ron

former IBM Field Engineering in mainframe era

-

[Carl Claunch]

Hi Ron

The poster uploaded it to Google Drive with permissions turned off for printing, downloading or other actions, which is why Google shows it in a reader that blocks all those actions. 

[David Kennedy]

Apparently not enough RTFM.
I think the permissions should allow download now

David

[Peter Diehr]

David and I had discussed complex variables - we certainly could have
done it, but:
(1) it would have taken more memory, and 8K was our limit for one of
the EMU computers,
(2) we were running low on time (David's graduation, loss of both IBM
1130's, my entry to grad school).
(3) the complex math libraries would require a lot of testing -
especially if we implemented the complex trig functions, etc.

So we didn't do it. It looks like David took it on later. Most of the
work was code generation, and the rest was following the template for
real variables - and tracing it through the symbol table. COMMON was
already a feature of the IBM Fortran compiler; the new variable types
had to be added to it. The name of a COMMON block becomes an external
reference, hence is processed by the CLB.

David must have implemented the Optize options; this is mostly
executable code generation. I don't recall ever discussing this
option, and we certainly didn't do it at EMU.

The logical variables, expressions, and the logical IF statement, were
the main new features - along with six letter variable names (IBM only
provided five). I figured out that "CLB" is the core load builder; we
had decided to leave this alone when we had looked at it earlier. The
changes to CLB are explained on P. 56 - the addition of a system
pseudo-register for complex numbers.

I don't recall creating a "double buffered card READ/Punch interrupt
service", we mostly read cards from hardware-buffered 2501 card
readers. I did write a multiply-buffered print routine for the 1403
printers; we had one for each computer. That was a lot of fun to write
(and debug), and it really sped up the printing. I also wrote the
interrupt handler for IRQ4 (Cancel Key), IRQ5 (STOP Key), and one for
the Single Step Mode, which was used for debugging nasty errors in the
system code - basically, it printed a running tab on the execution of
a program on the system console, with snapshots of the registers,
accumulator, etc. We never distributed this; I wrote it one weekend in
(probably) September 1971, because I was tired of cycling through the
single step mode for hours on end.

There is a mention (p. 79) of the CYTOS (Conversational Yale Terminal
Operating System); the IBM 1130 version was written by Tim Mellon (of
the Mellon family); I did the integration of the EMU Fortran with
CYTOS one weekend. I think it took about 40 hours. Gives one person a
conversational terminal as an editor for working with your Fortran
program. Since Tim Mellon owned his own IBM 1130, it was perfect for
him!

The documentation has been thoroughly revised; for example, the
references to the IBM manual page numbers is a change. Also the more
detailed examples. With a larger audience, better documentation is
worth the effort. Mine and David's efforts were subsumed, reformatted,
and improved. OTOH, they still used a deck of punch cards for the
documentation medium!

Best regards, Peter

> To view this discussion visit https://groups.google.com/d/msgid/ibm1130/cea8a7e6-015b-428b-aa4d-86f1da13a165n%40googlegroups.com.

[David Kennedy]

Thank you, Peter, that clears up a lot of the questions that arose.  So the EMU Fortran compiler and subroutine libraries are distinct and separate from the later DNA compiler, libraries, and DMS V2 replacement phases.  That explains the absence of the cited-in-the-DNA-manual CLB, DUP, and SUP phases from the card decks used to build the high school system that was once familiar to me.

Throwing a question to the group, has anyone expended the effort to build a SIMH disk image that includes the EMU compiler?
Warm regards
David