Tuesday, June 2, 2009

RIP: IBM Systems Journal (1962-2009)

This Spring, without Obituary Notice, Memorial Service, or 21-gun salute, IBM quietly interred the venerable IBM Systems Journal, thereby laying to rest one of the great artifacts of the company that built the modern computer industry and laid the foundations for our modern connected world. I served as Editor of the Journal from 1974 to 1977, the best job I had at IBM, so I’d like to say a few words.

Having seriously underestimated the market for computers in the 1950s, IBM entered the 1960s with a burning zeal to create a great company and a great industry. One barrier was education; it was becoming apparent that the optimal use of digital computers in the enormously varied situations in which they were to be applied would require significant knowledge of many areas and an over-all systems approach.

At the time, there were almost no systems science or computer science departments in universities. To fill this void, the IBM Systems Research Institute was established as a graduate-level academy for IBM employees. Shortly thereafter (1962) the IBM Systems Journal was founded as an adjunct to the institute to disseminate information about computing systems, programming, and applications. The great John McPherson, one of the true pioneers of our industry, served as the first Director of the institute and as the main advocate for the IBM Systems Journal (SJ).

As the industry grew, so did the number of overlapping professional journals, and by the early 1970’s IBM considered eliminating the SJ. When I became Editor in 1974, it was with the understanding that we were to continue refocusing the journal to what I came to call “advancing the current practice of computing,” while most journals chose to focus on theory.

The results were heartening. Circulation increased to 65,000 making the IBM SJ the most widely read technical journal in the computing industry. Despite our emphasis on practice, a study based on the Science Citation Index showed us to be the fourth most referenced journal. Other changes included efforts to put the journal closer to a paying basis by increasing the number of paid subscriptions and selling reprinted articles for use in classes and seminars.

But the real value of the journal was not its weight at the postage meter, but its effect on the computing industry.

A foundation element in modern software engineering is software inspections, consisting of a peer review of any work product by trained individuals who look for defects using a well defined process (see Wikipedia definition). The origin of this now-common practice was work done by Mike Fagin at the IBM laboratory in Endicott NY. Fagan had published a lab report describing inspections, which Associate Editor Al Davis found and presented to me. Despite a perceived lack of scientific rigor in the original report, Davis worked closely with Fagan to produce “Design and code inspections to reduce errors in program development, published in 1976. This paper is probably the most widely read and reprinted article ever published in any IBM journal.

Another effect on the industry was our role in promoting relational data bases. I tried very hard to interest Ted Codd, inventor of the relational model, to prepare a paper for us, but he preferred the professional society journals. We did, however, discover an IBM research project in San Jose named “System R” and we worked diligently with the research team to record these developments, publishing significant papers in 1977 and 1981. Larry Ellison read about System R in our journal and proceeded to develop the Oracle relational data base, thus transforming the IT industry and making himself (today) the fourth wealthiest man in the world.

So now, with the help of YouTube, it is time to organize our wake. I’ll play the Pogues’ “Body of an American.” You can join them singing “Danny Boy.” And perhaps we can all share A Wee Deoch-an-Doris for the IBM Systems Journal, for the many fine folks who contributed to it, and for Auld Lang Syne. (Warning –consuming this beverage would be cause for dismissal from the old IBM).

Footnote: for more detail, refer to ”A History of the IBM Systems Journal “ by George C. Stierhoff and Alfred G. Davis, IEEE Annals of the History of Computing, Vol. 20, No. 1, 1998.

No comments:

Post a Comment