Summary of the Stu Feldman Transcript (September 20, 1989)

Just as he brought a unique collection of skills to the UNIX project, Stu Feldman should bring a unique perspective to the UNIX oral history. It is important to recognize and respect his perspective on UNIX, because it could easily be lost in the fray. In his own words, Feldman was "peripheral to everything", and "always an oddball." Yet Feldman is surely selling himself short. Clearly a brilliant and multi-talented man (with a penchant for swallowing the occasional word), Feldman has much to tell us about UNIX. Hearing it directly from Feldman, it appears that he was peripheral not for any want of ability. Ironically, it may have been Feldman’s very versatility that left him feeling "peripheral." Feldman had to endure many project shifts within the larger UNIX effort. His impressive array of skills essentially forced him into the role of project mercenary, the fix-it-all man. Had he had a little more pure computing experience under his belt, such a profile might have earned him more of a managerial role.

Here we see another fascinating facet to Feldman’s resume. Of the major UNIX contributors, Feldman is the only Princeton undergraduate. He got his degree in astrophysics, with computing as "a side interest." It is important to note, however, that Feldman did develop a love for computers in the eighth grade. As he puts it, "I probably represent the first generation of computer brat." He went on to get his Ph.D in applied mathematics and theories of galaxies in 1973 at MIT(where else?). His obvious talents earned him work with Bell Labs as early as his sophomore year at Princeton. He began his work in Center 127, and Bell Labs thought confident enough in his prospects to install for his use what he believes was the first time-sharing terminal at Princeton. Despite the obvious attention he received from Bell Labs, Feldman maintains that he learned most of his computing informally, "in the gutter", as he puts it. He seems to have constructed his foundation of computing knowledge from an interesting variety of sources. These sources include summer jobs at Murray Hill, a couple of undergraduate courses, hacking in his spare time at the Princeton computer center, and one graduate seminar at MIT on automata. Here again, Feldman sees himself as a charter representative. This time he claims himself to be "one of the flowers of the (Murray Hill) culture." As far as his later work on UNIX is concerned, however, the most valuable component of Feldman’s computing experience was his work on Multics through Bell Labs.

The most immediate value of Feldman’s insights lies in his discussion of various technical aspects of UNIX. While it is difficult to pinpoint Feldman’s area of technical expertise, we can narrow down his central contributions to the UNIX project. Feldman’s work on f77 was integral to the project. Feldman’s main interests with respect to UNIX consisted largely of "compilers, programming languages, environments." In all likelihood, he wrote the first complete f77 compiler. He was able to use his extensive knowledge of the esoteric FORTRAN language to transform it into something more manageable, albeit still beyond the comprehension of the mere mortal. As Feldman explains, "EFL was a way of making FORTRAN look like C… I was basically gluing FORTRAN world onto the UNIX. FORTRAN had been used on UNIX, but it had never been inherent. That was actually one of my underlying goals… to make everything look portable and clean." This passage, like several other gems in the transcript, not only underscores Feldman’s value as a technical commentator, but also demonstrates the simple fact that Feldman’s most valuable passages are articulated very clearly. True to the tenets of oral history, Feldman speaks for himself, especially when dealing with the most important topics.

Feldman’s technical contributions include another project that was all Feldman’s own. He created make, a program that aided the efficiency of the UNIX project considerably. Feldman insists that make was "the smallest thing I did, but probably the thing I’m best known for." In fact, make as a contributor to UNIX is the perfect representative of Feldman: versatile, invaluable, and taken for granted.

Feldman clinches the indispensability of the transcript by offering much information about the internal workings of the UNIX project. Quite expectedly, Feldman has the definitive insider’s perspective. More than any other area, it is here that Feldman can give us unique observations. The very fact that he spent time working with so many different people on so many different projects within UNIX establishes him as an optimal spectator. Feldman paints a fairly rich picture of behind-the-scene life at UNIX. He portrays UNIX as an environment defined in large part by strange working hours, impromptu lunch discussions, show and tell's, controversial speaker presentations, and informal treaties.

To compound his technical and environmental knowledge of UNIX, Feldman also possesses an overview of the entire UNIX operation that is as intriguing as it is accurate. He even calls upon the ubiquitous car analogy, claiming UNIX to be the "MG" of early operating systems (an MG is a charming little sports car made by a garage-style operation in England). In short, it was a simple, elegant little performer lacking the speed and handling of a production sports car. More importantly, however, it never created a problem that it couldn’t fix on its own. Hear Feldman, "… I’ve been quoted more than once on the line that one of the great things about UNIX is that it lets you get out of the troubles it puts you into."

Feldman does himself a great service by describing himself as a sort of human glue near the conclusion of the interview. He glued FORTRAN onto UNIX, he helped glue broken UNIX projects, and in so doing, he indirectly helped to bring the various sectors of the project closer together. It would be hard to overstate this service. After all, UNIX itself has made a name for itself worldwide largely for its ability to connect a wide variety of users, bringing them closer together (as trite as it may sound). How can we overlook a man who helped unite UNIX, before it became a worldwide entity? We cannot forget the adage that the basic construction of a company’s product will largely mimic the structure of the company itself.

It would be going too far to say that the Feldman transcript could make or break a UNIX oral history. However, I do believe that Feldman’s transcript can serve a UNIX oral history in much the same way that Feldman himself served UNIX. Analyzed and placed correctly, Feldman’s transcript may just be able to fill in the gaps between the larger components of a UNIX oral history.

(Besides, how can we scrap his story? He was a Princeton undergraduate.)