Dusty Decks: Preserving historic software
by Paul McJones
Bob Bemer, who I communicated with recently regarding the Fortran “Tome”, died of cancer on June 22 [2004] at his home on Possum Kingdom Lake in Texas. He was the “father of ASCII” and one of the first to point out the “Y2K” problem (as far back as 1970). Many sources carried the story, including BBC News, The Washington Post, The Register, Reuters, and USA Today.
Update 2024/05/07: Added “[2004]”.
I learned that Peter Zilahy Ingerman, PhD, was the donor of the Fortran II listing at The Smithsonian. Peter published a number of books and papers in the area of programming languages and compilers.
I called Peter and had a very pleasant conversation. It turned out he’d donated the Fortran materials to The Smithsonian a number of years ago, and did not remember that specific item I was interested in, but he volunteered to travel from his home in New Jersey to Washington if that would help.
Peter is very interested in the history of computing, and has created a very detailed UNIVAC I/II emulator, a machine that Peter wrote software for between 1957 and 1963. Peter told me of a UNIVAC web page maintained by Allan Reiter.
Update 2016/1/2: Updated URL for Allan Reiter’s UNIVAC web page.
Update 2024/05/07: Updated URL for Peter’s emulator. Peter passed away on June 21, 2020.
I followed a link from Frank da Cruz‘s IBM 704 page at Columbia to George Michael’s 704 page, which is part of his Stories of the Development of Large Scale Scientific Computing. I sent an email asking if George or his colleagues remembered the Fortran ‘Tome’.
George replied:
1. I do not recall any book called the ‘Tome,’ although I was one of the first users of the original Fortran. We had four 704s but if the ‘Tome’ existed, I never saw it.
2. One of our programmers, Bob Hughes worked with the original Fortran development team. I interviewed him and the interview is posted on our web site http://www.computer-history.info/Page1.dir/pages/Hughes.htmlI will call him to check on the existence of the ‘Tome’ and if he (still) has a listing of the original Fortran. But I’m not overly hopeful. In any event, I’ll let you know what results I get.
Bob, who was listed as one of the authors of the original 1956 IBM Fortran manual, said this in his interview with George:
“I worked on what they called the “first-level” documentation. And I made the biggest mistake of my life by not bringing a copy of that home. Now you understand why I missed making my first million dollars.”
So I’m guessing he didn’t bring a copy of the ‘Tome’ back to LLNL.
Update 2016/1/2/ and 2016/12/29: Updated URL’s for George Michael‘s web site. He passed away in 2008.
Update 2024/05/08: Updated Columbia URL to https.
Earlier I had learned that Paul Pierce’s web site contains images of a number of system tapes for old IBM mainframes. Paul provides some utilities he wrote for making sense of these old tapes, which were written on 7-track drives, with 6-bit BCD characters and 36-bit words.
Tonight I came across Rob Storey’s emulation web site*. He’s written file utility programs BCD2IAS, IASMERGE, and FILEBROW that greatly simplify the process of reading, converting, and correcting the tape and card images from Paul Pierce’s web site.
I played with them for a while, and was able to view some of the Fortran IV compiler sources. More work will be necessary to create “fix records” for errors that BCD2IAS can’t detect by itself — it looks like there are occasional missing characters from blocked fixed-length lines, causing the source code to “drift” cyclicly to the left. But the result will be source code for the IBM 7090 Fortran IV compiler!
The bad news is that this compiler shares little if any source code from the original IBM 704/709 Fortran/Fortran II compiler (in particular, it did not contain the sophisticated code optimization components).
Rob’s web site also features an IBM 7094 emulator. (Paul Pierce’s web site has an IBM 709 emulator.)
* As of 2024/05/07 Rob’s emulation site is offline, but a snapshot is available from the Internet Archive Wayback Machine. There are also other versions; see https://github.com/Bertoid1311/B7094.
Update 7 Jan 2015: Updated URL for Rob Storey’s emulation web site.
Update 7 May 2024: Added footnote.
Every so often I try again to communicate with the Smithsonian. Eventually, I was able to make contact with Alicia Cutler, a Specialist in the Collection of Computers & Mathematics. She ran a search for me in their internal catalog, which produced a 12-page listing of documents related to Fortran. Three of them jumped out at me:
page 7, FORTRAN VOL I, 1959, “This is a FORTRAN program listing for the 704. Well documented.”
page 5, FORTRAN VOL II, “This is a FORTRAN program for the IBM 704, the title of which states “FORTRAN Editor Information”. Well documented”
page 5, FORTRAN VOL. III, 1959, “This is a FORTRAN program listing for the IBM 704. Well documented.”
I asked Alicia for more information on these three, and she replied, “I’ve looked over the documentation and have discovered that these three volumes relate to FORTRAN II, not FORTRAN. There is a letter from A.L. Harmon in May of 1959 stating that these are the SAP listings for the final FORTRAN II. It was split into three volumes but all together it is 1321 pages.”
Eureka!
J.A.N. Lee has had successive careers in civil engineering and computer science, and has been active in the history of computing for many years (see for example FORTRAN’s Twenty-Fifth Anniversary). He responded to my email saying, “I have asked about the original FTN compiler several times but without any success. Two sources would seem possible — John Backus himself … and the original recipient Westinghouse-Bettis.”
In a follow-on message he said, “I found two boxes of FTN materials relating to the 25th anniversary in a cupboard yesterday and will get them out to MV as soon as possible.
Is anyone doing anything about the fact that this is the 50th anniversary year for the start of the FTN development?”
As it happens, I’d recently looked through several boxes of Fortran materials J.A.N. had already donated to the Computer History Museum. They contained among other things photocopies of essentially all the papers included in his annotated Fortran bibliography (see FORTRAN’s Twenty-Fifth Anniversary).
Update 2024/05/07: Changed MW to MV [Mountain View?].