Remembering Carl Smith, 1950-2004

Sadly, Carl Smith passed away 10:30PM, July 21, 2004. He had had a 1.5 year battle with an aggressive brain tumor. He fought this battle with calm optimism, dignity, and grace. He is survived by his wife, Patricia, his son, Austin, and his sister, Karen Martin.

Carl was very active in the algorithmic or computational learning communities, especially in the inductive inference subarea which applies recursive function theory techniques.

I first met Carl when I interviewed for my faculty position at SUNY/Buffalo in the Spring of 1973. He was then a graduate student there and told me he was interested in recursive function theory. After I joined there, he naturally became my Ph.D. student, and that's when we both began working on inductive inference. We spent a lot of time together, pleasantly blurring the distinction between the relationships of friendship and advisor-student.

After Buffalo, Carl had faculty positions at Purdue and, then, the University of Maryland.

Carl had a very productive career. He was a master collaborator working with many teams around the world. Of course he also produced a number of papers about inductive inference by teams — as well as papers about anomalies, queries, memory limitation, procrastination, and measuring mind changes by counting down from notations for ordinals. I had the reaction to some of his papers of wishing I'd thought of the idea. This especially struck me with his 1989 TCS paper (with Angluin and Gasarch) in which it is elegantly shown that the learning of some classes of tasks can be done only sequentially after or in parallel with other classes.

Carl played a significant leadership role in theoretical computer science. In 1981, with the help of Paul Young, Carl organized the Workshop on Recursion Theoretic Aspects of Computer Science. This became the well known, continuing series of Computational Complexity conferences. Carl provided an improvement in general theoretical computer science funding level during his year as Theory Program Director at NSF. He was involved, in many cases from the beginning, in the COLT, AII, ALT, EuroCOLT, and DS conferences, as a presenter of papers, as a member of many of their program committees and, in some cases, steering committees. He spearheaded the development of COLT's Mark Fulk Award for best student papers and managed the finances.

Carl was very likable. He had a knack for finding funding to make good things happen. He was a good friend and colleague. He is missed.

August 2004 John Case