July 1990 - Vol. 33 No. 7

July 1990 issue cover image

Features

Opinion

Personal computing: simple complexity and COMDEX

One of today's emerging paradigms is the view that complex behavior or form can emerge from the interaction of relatively simple components, if you have enough of them and they have enough time to do whatever they do. The emergent behavior or form might seem systematic or chaotic. Some examples are neural nets, cellular automata, fractals, electronic mail networks, market economies, whirlpools, and snowflakes. Years ago, similar systems were often called self-organizing, and they were found in models of memory, pattern recognition, multilevel stores, and libraries. The area languished, however, awaiting the development of theory and powerful hardware. Personal workstations played an important role in facilitating experimentation and mass market personal computers are now up to the task.
Opinion

From Washington: budget FY 1991: the numbers tell the R&D story

This is the time of year when talk turns to fiscal budgets. In Washington, however, such banter typically involves astronomical sums of money.When President Bush released his proposed budget for FY 1991 last January, the reaction from the scientific community was mixed. Many observed that seldom have research and development (R&D) projects been given as prominent a place in a federal budget. Other industry watchers, while less enthusiastic, had to agree that in many respects R&D fared better under this year's budget than last year's.However, understanding the details of the budget is far more important than reviewing its broad outlines. For that reason, the American Association for the Advancement of Science (AAAS) calls members of the scientific community to Washington each spring to dissect, denounce and defend the government's R&D funding plans for the next fiscal year.The Colloquium on Science and Technology Policy conference centered around the analytical findings of the AAAS's Research and Development FY 1991 report. The three-day conference was peppered with high-ranking White House officials who either defined the specific branches of the government's R&D interests or discussed the possible implications the budget poses for future projects.There is an overall 7 percent budget increase for R&D, with a 12 percent increase in nondefense R&D programs and an 8 percent increase in basic research. In the area of computer science and engineering, DARPA, NSF, and the ONR remain the largest sources of government funds for R&D.Federal support in computer science is divided into two basic categories: defense and civilian. More than 60 percent of total federal R&D expenditures in computer science and engineering are supported by the defense sector. Moreover, federal R&D activities are conducted in government and nonuniversity labs as well as in universities. The majority of the funding for computer science research supports activities outside of universities.D. Allan Bromley, director of the Office of Science and Technology Policy (OSTP), explains the thinking behind the President's budget: To prioritize funding requests, the Office of Management and Budget follows three basic guidelines. They include 1. Programs that address national needs and national security concerns, 2. Basic research projects, particularly university-based, individual and small group research, and 3. Adequate funding for the nation's scientific infrastructure and facilities. Bromley points out that one of the primary avenues OSTP will emphasize this year is high-performance computing—a dynamic technology for industrial, research and national security applications. Of first concern will be the development of hardware to enhance mainframes and address the parallelism needed to make TERAOP computers perform trillions of operations per second. The next phase will be software development, followed by the construction of a fiber optic network.Bromley, who also serves as assistant to the President for Science and Technology, calls the FY 1991 budget an excellent one for R&D. However, he is quick to add there are problems with those numbers. (One of the most serious involves the funding rate for grants at the NSF and National Institute of Health (NIH). Despite a decade of funding increases, the money available for new, young investigators is very tight. Indeed, the scientist community is partly to blame, he says.“We argued for multiyear grants and contracts to cut down on the amount of paperwork required to do research,” recalls the OSTP director. “Both NSF and NIH have responded to those requests, and in the process they built substantial ‘outyear mortgages&rsquo’ for themselves.”
Research and Advances

Constraint logic programming languages

Constraint Logic Programming (CLP) is an extension of Logic Programming aimed at replacing the pattern matching mechanism of unification, as used in Prolog, by a more general operation called constraint satisfaction. This aritcle provides a panoramic view of the recent work done in designing and implementing CLP languages. It also presents a summary of their theoretical foundations, discusses implementation issues, compares the major CLP languages, and suggests directions for further work.
Research and Advances

The 20th annual ACM North American computer chess championship

Despite entering ranked almost a class above the field, a last-round loss forced DEEP THOUGHT to settle for a first-place tie with HITECH at the 20th Annual ACM North American Computer Chess Championship. The five-round Swiss-style tournament was held November 12-15 at Bally's-Reno in conjunction with Supercomputing '89. It marked the twentieth consecutive year that ACM has organized this major chess event. Until 1988, the tournament took place at the Annual ACM Conferences. In 1988 and again this year, however, the event was hosted by the joint ACM SIGARCH/IEEE Computer Society Supercomputing Conference. Ten teams participated in the strongest computer chess tournament in history. Every program was playing at least at the Expert level. This year's tournament offered $5000 in prizes. HITECH and DEEP THOUGHT's programmers each won $2000 for their first-place tie while MEPHISTO X and BEBE's programmers split the $1000 third-place prize. In addition to the cash prizes, trophies were awarded to the first three finishers. A special trophy was given to MEPHISTO X as the “Best Small Computing System.” A Technical Session chaired by Tony Marsland was held during the championship. The topic of the session was endgame play by computers. Once upon a time computers played the endgame particularly badly, but this is no longer the case. The session considered some of the improvements and some of the problems that remain. David Levy served as Tournament Director, returning after a layoff of almost a decade. He served as TD for the first time in 1971, continuing into the early 1980s when his own programs began to compete. Levy will take on DEEP THOUGHT in London in a four-game match in December.* In 1978, he won a bet made in 1968 that no computer would defeat him during the following ten years. This time he appears to be the underdog. Attending the championship as an Honored Guest was Ben Mittman. Mittman was head of Northwestern University's Vogelback Computing Center during the years that Slate, Atkin, and Gorlen's programs dominated the ACM events. Some give him credit for being Northwestern University's greatest and most successful “coach.” From 1971 through 1983, Ben also was involved in the organization of the tournaments, From 1977 through 1983, Ben served as the first president of the International Computer Chess Association. He was also the first editor of what is now called the ICCA Journal, the main journal for technical papers on computer chess. This year the championship is scheduled to be a part of Supercomputing '90 in New York City on November 11-14. The 1990 event will see the first major change in the tournament rules. For the last 20 years, the rules have specified that each player is given two hours to make the first 40 moves and an additional hour for each 20 moves thereafter. Games frequently lasted more than six hours. This year, each computer will be required to make all its moves in two hours, thus guaranteeing that no game will last more than four hours. In addition to the main championship, a special endgame tournament will be held testing the programs' abilities in this special part of the game. For the first time at Supercomputing '90, all games will be played during the day beginning at 1:OO p.m.—except for one 7:00 p.m. Sunday evening game on the 11th. The event will be a five-round Swiss-style tournament. For information contact Professor Monty Newborn, School of Computer Science, McGill University, 3480 University Street, Montreal, Quebec, Canada, H3A. 2A7.
Practice

Agenda: a personal information manager

The free-form, evolving, personal information that people deal with in the course of their daily activities requires more flexible data structures and data management systems than tabular data structures provide. A tool for managing personal information must conveniently handle freetextual data; allow for structure to evolve gracefully as the database grows; represent unnormalized data; and support data entry through database views. We have designed a new type of database that serves these needs—“item/category” database—and realized this design in a commercial personal computer software product named “Agenda.”

Recent Issues

  1. October 2024 CACM cover
    October 2024 Vol. 67 No. 10
  2. September 2024 CACM cover
    September 2024 Vol. 67 No. 9
  3. August 2024 CACM cover
    August 2024 Vol. 67 No. 8
  4. July 2024 CACM cover
    July 2024 Vol. 67 No. 7