STOREFRONT COMPUTING (p. 6). "The People's Computer Center, a not-for-profit corporation located in the San Francisco Bay area, is devoted to demonstrating and propagating the recreational and educational uses of computers. Its own literature bills it as 'a loosely-knit community of students, educators, programmers, technicians and gamesters' which currently possesses two computers, six teletype terminals, a video terminal, a plotter, and some hand calculators, plus math activity materials, puzzles, and board games. These resources are used for school field trips in which students as young as 5 years old learn to play computer games, for classes in BASIC and math labs, and for teacher workshops and informal open-house sessions."
COMMUNITY COMPUTING (p. 7). "… the Community Memory Project [is] a sort of local public access information network begun last August in Berkeley under the auspices of a non-profit collective known as Resource One.
"Community Memory began with a single teletype terminal placed in the lobby of Leopold's Records in Berkeley, which people were encouraged to use as a combined electronic bulletin board and data store. Since then thousands of people have discovered the terminal and typed in messages, classifying their items themselves so other people can find them quickly. The collection, with over a thousand active items now, includes exchanges traditional to other public media—bulletin boards, classified ads, etc.—as well as exchanges and dialogs which are developing their own unique forms. There are cars for sale, rock bands looking for bass players, carpenters looking for jobs, groups offering counseling, tennis players looking for partners, political commentaries, etc.
"Other public access terminals have since been installed, and the Community Memory is beginning to grow into a network."
KEYNOTE ADDRESS (p. 10). "Before a standing-room-only audience of over 600 at the Fall meeting of the DEC Users Society in San Diego last month, Captain Grace Murray Hopper delivered a keynote address … on the topic, 'Possible Futures, Hardware and Software,' [in which] she directed her remarks toward the information processing community's need to react ever more quickly to the future—a future which, she said, is approaching more and more rapidly each day.
"'In the oncoming world of shortages, managers are going to need more and more information—i.e., processed data—faster and faster. And that means more computers,' she said.
"But in Captain Hopper's mind, more computers doesn't mean bigger computers. Future systems, she predicted, will be bigger systems of small computers. Such distributed systems are made possible by the rapidly falling price of hardware."
"Captain Hopper called for more and better training, as well as more sharing of ideas among computer professionals, including active support for professional associations."
DIGITAL SYSTEMS EDUCATION (p. 15). "The National Science Foundation has awarded a two-year grant to establish a project which has been christened the DISE (Digital Systems Education) Committee. Project organizers are Dr. J. T. Cain, Dr. R. G. Hoelzeman, and Dr. T. W. Sze of the University of Pittsburgh, Dept. of Electrical Engineering.
"'For some years now, there has been a very wide range of differences in the methodology and philosophy of digital systems education,' Dr. Hoelzeman explained. 'The differences range from solely software oriented type programs to solely hardware-development type programs. Even programs with the same hardware-software mix have great differences in implementations and directions.'
"'These observations have led to the belief that there is now an urgent need for a permanent inter-university, inter-industry working group to coordinate, develop and distribute digital systems educational technology and materials,' Dr. Hoelzeman said. 'The purpose of the NSF grant is to provide the catalyst to institute such a group.'"
HARDWARE DESCRIPTION LANGUAGES (pp. 18–19). "Likewise, computer hardware description languages, which are nothing but another application of symbology, are essential to computer design. The use of computer languages to describe digital system designs can be traced back to Shannon's work on switching circuits in 1939, Aiken's work on switching theory at Harvard in the 1940's, the logic diagrams at MIT and the National Bureau of Standards in the late 1940's, the flipflop equations in the 1950's, and the register languages in the 1960's. Interest in this subject has grown over the past decade and a number of languages have been developed (e.g., AHPL, CDL, DDL, ISP, and PMS) to describe computer structures and hardware algorithms. Some of these languages are so created that one does not need to know electronics. In the case of others, simulators are available for use in experimentation and design checkout."
PATROL CAR COMPUTER (p. 73). "When city and state police at Wayne, Mich., raided a junkyard believed to be the [stolen car] ring's headquarters, they were confronted with the task of laboriously checking vehicle identification numbers of more than 500 automobiles to find stolen cars among them.
"… Wayne police called for help from the nearby Dearborn Heights police department. Each of the 15 Dearborn Heights patrol cars carries a newly installed two-way portable computer, the Arcom MCT-16, manufactured by Atlantic Research Corp. of Alexandria, Va.
"As workmen pulled each car from the junkyard stacks, a policeman relayed the car's vehicle identification number to a Dearborn Heights patrol car stationed in the junkyard. …"
"More than a dozen stolen vehicles were found among the first 80 cars checked out."
PARALLEL PARADIGMS (p. 38). "Parallel computers have failed to make a major impact on mainstream computation, despite the fact that commercial products have been available for almost a decade. A substantial performance/price advantage over conventional supercomputers, and even over large uniprocessors, has not been enough to convince users to move from a sequential to a parallel mode of computation.
"An examination of the state of the art in parallel computing suggests an explanation. Different classes of parallel architectures require radically different paradigms for describing and executing computations. In addition, both practitioners and theoreticians have specialized along architectural lines. There is no obvious winner among these architectures; it is hard to move applications from one class to another; and many potential users are unwilling, on the present evidence, to make a computer acquisition decision with long-term implications."
APPLICATION PHILOSOPHIES (p. 52). "In recent years, attention has focused on software engineering—especially its computer-aided aspects—as a 'cure' for the life-cycle problems of complex computer-based systems. Unfortunately, this emphasis is like placing the cart before the horse. Software engineering tools and methods are important, but they should be the natural result of a well-developed philosophy for solving the application problem.
"By philosophy I mean a unifying common view of how a problem or class of problems shall 'in principle' be treated. The view, which is based on concepts, must be commonly held by project team members and all other parties with vested interests. It involves the development of a strategy from which decisions (large and small) emanate. The philosophy should be documented, most often via appropriate paradigms and models, for communication to others. Once the philosophy is understood and practiced, the decisions will follow a common pattern."
BENCHMARKS (p. 74). "Advice for users looking at benchmark numbers to characterize machine performance should begin with a warning not to trust MIPS numbers unless their derivation is clearly explained. Here are some other things to watch for:
• Check whether Mflops numbers relate to a standard benchmark. Does this benchmark match your applications?
• Know the properties of the benchmarks whose results are advertised.
• Be sure you know all the relevant facts about your system and the manufacturer's benchmarking system. For hardware this includes clock frequency, memory latency, and cache size; for software it includes programming language, code size, data size, compiler version, compiler options, and runtime library.
• Check benchmark code listings to make sure apples are compared with apples and that no illegal optimizations are applied.
• Ask for a well-written performance report. Good companies provide all relevant details."
OPEN SOFTWARE (p. 96). "The Open Software Foundation has announced the first release of the OSF/1 operating system. It is compatible with Unix System V and Berkeley programming interfaces and can be used in the Intel 302, Digital Equipment Corp.'s Decstation 3100 workstation, and the Encore Multimax multiprocessor system, among others.
"OSF/1 supports the OSF/Motif graphical user interface and employs a Mach-based kernel from Carnegie Mellon University that allows workloads to be distributed among multiple processors. The kernel offers dynamic system configuration, logical volume management, and disk mirroring."
"OSF plans to release subsequent versions of the system every 12 to 18 months, culminating in microkernel implementations."
ROBOTS (p. 112). "Industrial robots have undergone many developments since the arrival of the first Unimation machine in 1962. Applications are widespread and growing. The robot's advantages of flexibility, reprogrammability, tirelessness, and hardiness have come to be appreciated by industrialists. Even the layman realizes that today's industrial robot, unlike the tin marvel of science fiction, has a real and useful role to fulfill.
"However, despite many developments in the associated technologies, the industrial robot capable of sensing and reacting to the external environment is still in its infancy. Robots have been and continue to be shaped in the image of man, and the second generation is seen to be a further step in this direction. However, anthropomorphism is a constraining influence, and new perspectives are needed. Perhaps we should endeavor to see beyond the principles of replacing a human by a human-like robot."
PDFs of the articles and departments from Computer's December 1990 issue are available at www.computer.org/computer.