Thursday, October 9, 2008
JonL White will speak at Lisp50
White (a.k.a. "JonL") was first instructed in the way of computers--of numeric and symbolic computation--whilst in High School at a special "University" school associated with the Illinois state university system. The year, coincidentally, was 1958. He and a fellow mathematics student were subsequently hired by the University Computer Lab, where they jointly wrote code for the IBM 650, and actually developed a working compiler for next machine, the IBM1620. JonL received a Bachelor's degree in Mathematics from Carnegie-Mellon University, and a Masters in Applied Mathematics from Harvard University; after he had completed all the requirements of the Ph.D. program (ABD), he took a leave of absence to work at the MIT AI Lab in the fall of 1969. During the next few years, he advanced the state of Lisp development in many ways, and supervised numerous part-time undergraduates who assisted in the work, such as: Hillary Orman, Richard Stallman, Guy Steele, Eric Rosen, and Stavros Macrakis.
During the year 1977 he took a sabbatical leave to work at IBM's Watson Research Laboratories on the Lisp/370 project there, and returned to MIT the next year to head up the NIL project (NIL: a "New Implementation of Lisp.") By early 1980, he had distributed PDP-10 Maclisp to about five dozen other University sites (most with "AI" research groups.) After a meeting in April 1980 with the government agency mainly involved in AI research funding, JonL, Dick Gabriel, Scott Fahlman, and Guy Steele (by then at CMU, and involved with Gabriel at LLL) observed that the aforesaid four were involved in at least four different Lisp systems going in "five different directions"; in short order, they instituted a movement to coalesce efforts into a common direction. A bit later, after Dave Moon joined the conversations, the effort got a name: "Common Lisp."
JonL has played a signal part in the Association of Lisp Users, from the second conference hosted by this group -- the Lisp Users and Vendors Conference, 1991 in San Diego where he was the conference chair. Later he was elected to the Board of Directors of the ALU, and served as Program Chair for the International Lisp Conferences 2005 and 2007. He also played a significant part in the ANS Standardisation of Common Lisp, taking place between 1987-1994; he authored the LOOP proposal, and contributed to the Compilation and CLOS subcommittees.
From Massively Monster Machines to MicroChips: Forces Affecting Lisp Language Design through Five Decades
This talk will explore some of the constraints imposed by the character of computer hardware on Lisp's design over the past half-century, as well as being shaped by research goals. It will be presented mostly as an account by someone who was both an eyewitness and a protagonist in the struggle to create a better programming language. Even his own background as an undergraduate and graduate student (at CMU and Harvard respectively) seems to be relevant: for example, frustration with the use of IPL-V in AI classes at CMU, with the use of COMIT at Harvard to code up an Algebraic Simplification project (having no Lisp then!) And interestingly, a major influence of metamathematics shows up in his work on Lisp, giving rise to the notion that "Numbers are Symbols Too." A strong graduate education in mathematics had substantial influences on the ways he approached most significant issues in the use of, and the design of, Lisp -- in directions that often differed from his contemporaries. More specific topics to be discussed in a historical context: the unique, long-lasting alterations and additions to early lisp he made at MIT in the late 1960's and early 1970's time frame, the standardisation process for ANS Lisp, and the evolution of Lisp vendors during the two decades beginning in 1980.