close this bookVolume 2: No. 27
View the documentNews -- research support
View the documentNews -- NSF program deadlines
View the documentThe following are from the NSF Bulletin
View the documentNews -- supercomputing; HPCC/NREN
View the documentNews -- research employment
View the documentResources -- career services
View the documentNews -- job opportunities
View the documentNews -- publication opportunities
View the documentNews -- legal issues; content-addressable memory

The House VA, HUD, and Independent Agencies Appropriations Subcommittee has made its recommendations on NSF funding for FY'93. (The House allocates budget levels for agencies and programs, then considers them in new groupings for final appropriation and authorization. NSF competes with the Veterans' Administration, Housing and Urban Development, NASA, EPA, and several independent agencies. A bit of bad press such as the recent NSF behavioral-science rescission can shift large sums from one agency to another.) The subcommittee had $3B more than last year, most of which went to the VA and HUD. The Space Station was cut to $525M below request. NSF was recommended at $2.723B, just $150M more than last year. Salaries and Expenses will be cut by $19.5M. (Warning: You may get the level of service that you pay for!) NSF's Critical Technologies Institute should get its $1M budget request. Full Committee mark up in the House may not occur until mid July. [, grants, 6/29.]

The Commerce Dept. has announced 27 Advanced Technology Program awards to 80 institutions. About $90M will be given in this second year's competition, with another $100M planned for later years of continuing awards. Eleven projects involve electronics, data storage, and semiconductor technology, including an HDTV fractal compression chip from Iterated Systems (Norcross, GA). Other awards are mainly in materials and manufacturing. Individual-company awards are limited to $2M of direct R&D costs over three years. [Brian Robinson, EE Times, 5/4.] AI systems in support of manufacturing might be a way to tap this program. The National Center for Manufacturing Sciences will get $19.7M over five years. Honeywell, Hercules Aerospace, Sheldahl, and 3M will get $2.3M over three years to develop and apply a generic neural-network control system. [Joanne Connelly, Electronic News, 5/4.]

Congress and the President are still deeply divided over industrial policy. Last year, Congress required the administration's National Critical Technologies Panel (under Alan Bromley) to put forth a list of critical technologies. Debate over the report continues, but both sides acknowledge that critical technologies exist. A new advisory panel called the Critical Technologies Institute (CTI) appears to be the next forum with bipartisan credibility. CTI might spark action even before the election. [Brian Robinson, EE Times, 5/4.] Critical technologies will get most new government grants, of course.

Sematech's William J. Spencer advocates that the 10 major national labs be converted to research houses working in cooperation with industry consortia (such as Sematech). Lab personnel should not be expected to develop and commercialize their own products. Working through consortia would provide project prioritization and guidance without unduly benefiting any one company. [Electronic News, 4/27.] (The more efficient that national-lab tech transfer becomes, though, the more difficult it will be for small businesses to compete.)

Roger Woolnough says that Esprit has resulted in very few commercial products, despite thousands of teams collaborating on hundreds of projects. Major European companies (Philips, Olivetti, Thomson, Nixdorf) have not been regenerated; rather, they have suffered financial setbacks. The European Commission now wants to see "priority technology projects" and producer-user collaborations (as opposed to precompetitive research). Common Market countries may be unwilling to fund the requested budget increases, though; the commission is "somewhat in the position of a wastrel son, promising to reform if only his father will clear his gambling debts one last time." [EE Times, 5/4.]

The Office of Technology Assessment has estimated that 127K defense engineering jobs will disappear by 1995. [IEEE. EE Times, 4/27.] That means serious competition for remaining slots in all engineering sectors.

NSF has been widely criticized for spreading (or at least permitting) the impression that cumulative "enrollment shortfalls" -- under an arbitrary highest-enrollment-to-date line -- foreshadowed a serious shortage of scientists and engineers. Bob Bellinger faults former NSF Director Erich Bloch for focusing on input to the scientific pipeline instead of the productivity of mid-career and older engineers. Bob does like the text of a recent protest from Bloch, though: "If the present recession, our miserable productivity growth, and our continuing loss in market share are to be reversed, we need additional incentives and encouragement for scientists and engineers. ... Members of Congress, lawyers and MBAs are not going to improve our competitive standing in the world. Scientist and engineers just might." [EE Times, 4/27.]

(I find it difficult to hold competitiveness and NSF in the same thought. Economic contribution may be considered -- to a small extent -- for SBIR proposals, for large engineering centers, and perhaps for a few other engineering awards. For most proposals, though, scientific merit is the touchstone. Review is by scientists, not industrialists. I have hopes the Commerce Dept.'s Advanced Technology Program will maintain a more pragmatic focus.)

There are three stages of R&D-lab evolution, according to Tamara Erickson, a VP at Arthur D. Little Inc. The first generation is a beautiful R&D center with few links to corporate offices, funded by a fixed percentage of sales or revenue. Projects are chosen without administrative interference. The second generation is adversarial, with "eggheads" begging for funds from "pencil pushers" who keep killing bottom-ranked projects until little is left. Researchers are afraid to abandon false starts because they'd be terminated, and they discourage realistic managerial oversight for the same reason. The third generation is cooperative, and is mainly seen in new biotechnology firms. [Robert Bellinger, EE Times, 5/4.] Partnership only works when each side needs the other. Second-generation researchers are in danger because they have failed to demonstrate the need for their services.

Bill Park ( sent me extracts from a recent "glut" discussion. Joel Norris (joel says that we just have fewer trained people than we need and more than the country is willing to pay for. attributes short-term R&D management to merger mania (caused by the Reagan administration's refusal to enforce antitrust laws). Any company with enough funds for fundamental research became a target. Don Gillies (gillies notes that U.S. research hiring has seldom been in equilibrium: We've had the space race, AT&T breakup, government-supported PhD overproduction, undergraduate baby boom, government-unsupported research overproduction, 1992 faculty glut, and now recession-induced continuing enrollments. We have yet to reach steady state in terms of CS retirements. Ellen R. Spertus ( advises keeping an eye on research fashions. Build your expertise in an area that departments care about. Computer security is out. Sequential is out; parallel is in. Brian Yamauchi ( expects genetic algorithms and artificial life to become mainstream in the next couple of years, but to fade quickly as all AI approaches have. Eugene Miya ( says "do your own thing." He remembers some mid-70s astronomy graduates who couldn't get top jobs, so they formed their own "observatory" and begged for equipment. Now they have an observatory and their own institute.

Maria Zemankova ( has a description of the CISE/IRIS initiative for Research on Scientific Databases (NSF 92-65). Three research areas are particularly encouraged: scientific database models and systems; knowledge discovery in scientific databases; and advances in resource sharing environments. Write to Maria or call (202) 357-9570 for an online announcement. Proposals are due 7/15, and may request $300K/year for three years. $2M-$3M will be awarded this year.

CISE Instrumentation (NSF 91-52), 8/3. Multiproject, special- purpose CS equipment or instrumentation. Caroline Wardle, (202) 357-7349.

CISE Institutional Infrastructure Program (NSF 91-7), 9/21. John Cherniavsky, Cross-Disciplinary Activities, (202) 357-7349.

NSFNET Program (NSF 90-7), 7/1. Daniel VanBelleghem (dvanbell, Networking and Communications Research and Infrastructure, (202) 357-9717.

U.S.-Taiwan Cooperative Science Program, 7/15; William Chang, International Programs, (202) 653-5343.

Human Cognition and Perception (NSF 88-62), 7/15. Joseph Young, Behavioral and Cognitive Sciences, (202) 357-9898).

Linguistics (NSF 88-62), 7/15. Paul Chapin, Behavioral and Cognitive Sciences, (202) 357-7696.

Decision, Risk, and Management Science (NSF 91-59), 8/15. Robin Cantor or John Castellan, Social and Economic Science, (202) 357-7569.

Law and Social Science (NSF 89-9), 8/15. Michael Musheno, Social and Economic Science (202) 357-7567.

Methodology, Measurement, and Statistics in Social Sciences (NSF 89-9), 8/15. James Blackman, Social and Economic Science, (202) 357-7966.

Long-term Research Opportunities in Japan (NSF 90-144), 9/1. U.S. researchers can visit Japanese government, university, and industrial laboratories for six to 24 months.

Nominations for the 18th Alan T. Waterman Award for outstanding young scientists, mathematicians, and engineers are due by 12/31. Susan Fannoney, National Science Board, (202) 357-7512.

A new ruling by the Commerce Dept. restricts export of supercomputers with Composite Theoretical Performance greater than 195 million theoretical operations per second (Mtops). Different levels of licensing will be required for different countries. A Cray 1 from the mid-1970s would not be exportable to the Middle East, for instance. Mini-supercomputers and even some workstations are above the threshold, and there is currently no indexing provision to change the definition of "supercomputer" as technology changes. Cray Research is not happy with the definition, but is glad that it now has a definite rule to work with. [Brian Robinson, EE Times, 5/25.]

Dave Hughes ( has expressed a bit of concern over the politics of the National Research and Education Network (NREN): "Everything is in Virtual Unreality, where laws don't seem to apply, intents of Congress are irrelevant, the text of passed bills are Rorshach Ink Blots, cross-membership in one another's deals or organizations are so convoluted as to appear incestuous. And those who ought to be speaking out the loudest against NREN being implemented wrong are the quietest for fear they won't be in good favor by one winner-loser faction or another when it is implemented." [com-priv, 6/20.]

Gordon Cook ( is wondering who are the scientists who need the high-bandwidth communication lines that Congress and NSF propose to provide. Godon worked for three years at the JVNC Supercomputer Center. He says you can use a 2400bps modem to connect to "any Cray out there" in four NSF-sponsored centers and 38 state supercomputer centers. Output on a tape cartridge can be delivered the next day by Federal Express. If computations must be steered interactively, you just have to travel to one of the centers. [com-priv, 6/20.]

Massachusetts has the world's highest concentration of high- performance computing companies, including Alliant, Convex, DEC, Kendall Square Research Corp., Meiko Corp., Thinking Machines, and Wavetracer Inc. Massachusetts universities turn out 12% of the nation's PhDs in computer engineering, 10% in mathematics, and 7% in EE. State groups are looking to government HPC funding ($6B over 5 years) and subsequent sales to improve the job market. Revenue in massively parallel processing is expected to jump 90% this year and 25% per year for several more years. The 1,500 software companies in the state will also benefit. [Margaret Ryan, EE Times, 5/25.] (Time to apply for a job?)

DEC will build a research and manufacturing facility in Hudson, MA, for continued development of its Alpha RISC architecture. [CW, 6/8.] Meanwhile, only 3,700 workers took advantage of DEC's early retirement offer (including William Heffner, VP of image, voice, and video applications). Layoffs will be used to reach a 10,000 reduction this year. [Melinda- Carol Ballou, CW, 6/8.]

Intel's Touchstone Delta, with 513 I860 processors, is one of the most powerful computers ever built. It's tricky to program, though. Programmers at JPL have found that ported C and Fortran needs a high degree of custom programming and tuning. [Jean S. Bozman, CW, 6/8.]

The Federal Home Loan Mortgage Corp. (Freddie Mac) has ordered a Cray Y-MP EL, the first Cray to be used for financial service analyses. [CW, 6/8.]

NCR will build a $30M research and manufacturing facility in Rancho Bernardo, CA. Development will focus on massively parallel computers, including database technology acquired from Teradata Corp. (through parent AT&T). [CW, 6/8.]

I'm a couple of months behind on this, but BBN's Systems and Technologies Division (Cambridge, MA) has had R&D openings in sonar surveillance and in speech and network technologies. Telecommunications jobs in the Northeast are opening up, including an expected 1,000 at BDM International Inc. (Cambridge, MA).

Hiring on the West Coast should pick up near the end of the year, but companies are still looking for software engineers and senior technical people with five years experience. Hot areas are C/Unix, C++, object-oriented anything, GUI and windowing environments, and OS/2. [David Pregeant, Source Engineering, and Trish Murphy, Source EDP. Robert H. Blissmer, EE Times, 5/25.]

Starting salaries are still rising steadily, although BSEE graduates below the top 10% are having trouble finding jobs. Offers average $34.1K, up about 2.8% since September. Many EEs are heading for graduate school and teaching positions. (Imagine the crunch at the "pyramid top" when all of them graduate!) Others are looking to consulting companies, small businesses, and programming jobs. Chemical, mechanical and civil engineers at RPI are getting more offers and better starting salaries, with petroleum engineers starting at $41K. Business graduates typically get $24.3K. [Robert Bellinger, EE Times, 5/4.]

Business Week (6/29) lists 1991 R&D expenditures by U.S. companies (excluding R&D contracts). Big spenders include GM ($5.8B), IBM, Ford, AT&T, DEC, Kodak, HP, Boeing, GE, and du Pont ($1.3B). If you prefer a company devoted to R&D, consider Software Publishing, Chips & Technologies, MIPS, or On-Line Software International. (Biotech companies have even higher expenses per employee.) Or look for growing R&D labs, such as NEC's $32M complex in Princeton, NJ. Bell Lab's John Mayo says that the new emphasis must be on research AND development, with integrated efforts leading to timely products. BW's table lists 41 companies in software and services, but the interesting jobs might be in other high-R&D companies. (Medical instruments, for instance, or aerospace.)

David Pregeant of Source Engineering (Palo Alto) says that the Silicon Valley professionals most in demand are software engineers experienced with Windows, OOP, C++, real-time embedded control software, or graphics applications. [Kate Colborn, EDN News, 4/30.]

According to Francis K. Walnut, software experts from engineering, information science, and CS share an inability to deliver working systems. Software engineers and (especially) management information graduates are taught that the interesting, creative, important part of software development is over when an architecture is chosen and specs have been written. Implementation is just an unfortunate necessity. Computer scientists imitate natural scientists seeking basic principles and fundamental truths, and believe that any effort lacking in scientific rigor is necessarily inefficient. Walnut says that real productivity comes from experienced, end-to-end programmers -- often working on computers in their basements. [CW, 6/15.] (Going a step further, a programmer is likely to take the greatest care if he or she will be responsible for documentation, training, maintenance, and bug fixes. Apple's one-button mouse and graphic interface, for instance, came from programmers striving for simple documentation and ease of use. Call in experts when necessary, but to keep responsibility from diffusing. It's easiest if you own your own business.)

Edward Yourdon takes the opposite view in "Decline and Fall of the American Programmer." He says that American programming shows sloppiness, sloth, mediocrity, and lack of discipline. Our only hope against foreign competitors lies in modern tools, structures, techniques, methodologies, and management. [Electronic News, 5/18.] Alan Zeichick agrees that AI programmers should know about the software methodologies Ed discusses. There's more to software development than just algorithms. Alan thinks that the quality problem is being addressed, though, as evidenced by lengthy beta tests and missed shipping dates. [AI Expert, 7/92.]

NSF has a Vacancy Hotline listing all current vacancies: (202) 357-7735 or (800) 628-1487. Salaries are $32.4K-$60K for assistant program directors to $54.6K to $85.5K for a program director. Temporary appointments require only a resume and current salary; permanent positions require a Form 171. [NSF Bulletin, 6/25.]

Some 700 national employers have job hotlines, including industry, government, universities, and military facilities. "Job Hotlines USA" is a new 100-page paperback listing. $16.95 plus $3 S&H from Career Communications (Harleysville, PA), (800) 942-2417. [SJM, 6/28.] (Most publishers pay the S&H when you order direct, since you save them a dealer discount of up to 40%. You might try negotiating.)

"Technology Jobs Guide" lists 600 high-tech San Diego companies. $24.95 from San Diego Technical Books (Kearny Mesa, CA); (619) 279-4990. [EDN News, 5/14.]

Jobhunt is a PC program that manipulates addresses and profiles of CS and engineering companies. It can search, select, sort, edit letters, merge contact information, and print labels. $49.95 from Scope International (Charlotte, NC), (704) 535-0614. [EDN News, 4/30.]

The recession has spurred professional societies to work harder and provide more service. (Funny how that works. An inverse correlation between funding and personal service.) Videotapes and training courses are becoming common. The Bay Area chapter of the Assoc. for Women, (415) 905-4663, has added a job-seekers SIG, bboard service, bimonthly lead-sharing meetings, and a detailed membership directory. The Boston Computer Society, (617) 252-0600, is providing employment counseling and resume-writing help, as well as discounts on outside products and services. The Corporate Assoc. for Microcomputer Professionals, (708) 291-1360, offers trade-show and conference admissions. Two other groups are the Data Processing Management Assoc., (708) 825-8124, and the Tele- Communications Assoc., (818) 859-2652. Volunteer work within such an organization can provide valuable leadership training. [Suzanne Weixel, CW, 6/8.]

Discouraged job-seekers might enjoy "E-Mail Links Science's Young and Frustrated" in Science (5/1, p. 606). [Tony Lai (, Bill Park.]

Bill Clancey is resigning as Editor of the AAAI Press, and his replacement is being sought. A series of symposium technical reports is planned, starting with reprints and moving into electronic publishing. [AI Magazine, Summer '92.]

Teleos Research (Palo Alto, CA) needs an experienced MSCS as systems admin (Unix workstations, Macs, and Lisp Machines) and research programmer in real-time binocular machine vision for goal-directed robotic action. Ann Reid (, (415) 328-8800.

UWestern Australia's Robotic Sheep Shearing Project needs an R&D engineer for robotic development and testing. Mr. E. Tabb, +61 9 310 7500, +61 9 310 7511 Fax. [James Trevelyan (james, comp.robotics, 6/19.]

Rockwell Space Division (Los Angeles, CA) has an MS/PhD opening in robotic effector R&D. Kenneth Goldberg (, comp.robotics, 6/22.]

A Washington, DC, company needs an experienced AI/KBS software engineer for Unix/X Windows/OOP development. ASC Software, 205 Dale Drive, Silver Spring, MD 20910. [Adam Joseph (,, 6/18.]

NIMH (Bethesda, MD) has an AI postdoc opening in cognitive neuroscience. Real-time data acquisition and process control experience would be helpful. Roger Erickson ( [, 6/24.]

Universiti Sains Malaysia needs a CS professor with 10 years experience in expert systems, NLP, parallel programming, networking, information systems, or related field. Appointment for 1-6 years, 3 preferred; salary to US$60K plus return passage, gratuity. Zaharin Yusoff, CS, 11800 USM, Penang, Malaysia. [,, 6/25.]

The AI Center at SRI International (Menlo Park, CA) needs a Motif/C/Lisp graphics programmer for a DARPA-sponsored project in terrain visualization. Additional projects available after 15 months. An HCI consultant may also be needed. Yvan Leclerc ( [ba.j.o, 6/27.]

The Centre for Computers in Law and Finance, Brunel University (Uxbridge, UK), has research openings in OODB, IR, GUIs, document management, and document imaging. A research fellow could earn up to #25,781. Vijay Mital ( [m.j.o, 6/26.]

Oryx Press needs 50 reviewers knowledgeable about women's educational equity and/or educational publishing. You would do up to three reviews per year for three years, with an honorarium for each review. Send resume by 7/1 to Tracy Moore, The Oryx Press, 4041 North Central at Indian School Road, Phoenix, AZ 85012-3397; (602) 265-6250 Fax. [arhjb@asuacad.bitnet, m.j.o, 6/22.] (Do AI publishers pay reviewers?)

URochester has predoctoral and postdoctoral openings in biostatistics (e.g., toxicological studies). Martin Tanner ( [m.j.o, 6/24.]

Pencom Systems (New York, NY) needs Smalltalk/C++ programmers for a major investment bank. Jaime Fields (jaime@pencom.uucp), (212) 513-7777. [m.j.o, 6/24.]

The Space Telescope Science Institute (Baltimore, MD) want an experienced MS knowledge engineer to continue development of a Common LISP scheduling system. Glenn Miller (, or call Robert Teleky at (410) 338-4882. [m.j.o, 6/25.]

The Lister Hill National Center for Biomedical Communications, Information Technology Branch (Washington, DC), needs C++ developers for a state-of-the-art information retrieval system. Charles M. Goldstein ( [m.j.o, 6/29.]

United Parcel Service R&D (Danbury, CT) needs Unix/C/Sun software engineering consultants experienced in image processing, AI, and OCR. Dr. Michael Moed (, UPS R&D, 51-53 Kenosia Ave., Danbury, CT 06810; (203) 731-6303. No email resumes. [m.j.o, 6/29.]

Elaine Rich wrote a nice editorial in the Summer AI Magazine. I particularly like her observation that "Occasionally, because we came from a very rich family (computer technology) with noble ancestors (the philosophy of mind), a story about our work would arouse interest in the wider community." She and the publications committee are soliciting articles of interest to the AI community as a whole.

AAAI is giving me a press pass for the big AAAI-92/IAAI-92 conference in San Jose this July 12-16. (Avron Barr suggest that I ask for one. Thanks.) You might get passes to other conferences if you contract to write reviews for publications. Advance plugs (like this one?) are also good leverage.

Intelligent simulation for high-autonomy systems; Int. J. in Computer Simulation, 1993. Contact Albert Y. Zomaya (zomaya or George W. Zobrist ( by 7/10. [NL-KR, 6/24.]

Processing natural language; IEEE Expert special track. Contact Terry Patten ( by 10/1. [NL-KR, 6/24.]

Neural networks for vision and image processing; The Asia Pacific Engineering Journal Part A. Contact Dr. V. Srinivasan ( or Dr. S.H. Ong (eleongsh@nusvm.bitnet) by 12/31. [Neuron Digest. 6/25.]

Analog VLSI computations (including speech/visual processing and artificial neural network learning); J. of VLSI Signal Processing. Contact Michael D. Godfrey ( by 10/1. [John Lazzaro (, connectionists, 6/26.

Presence: Teleoperators and Virtual Environments is a new quarterly from MIT Press Journals. Rebecca McLeod, (617) 253- 2886. [AI Magazine, Summer '92.]

Interesting name: There is (or was) a Computist Magazine and BBS for Apple II and IBM PC code. About $24 U.S. from 33821 E. Orville Road, Eatonville, WA 98328. [Robert W. Igo,, 6/20. Bob Cowles (]

Josef Skrzypek ( is collecting material for a book on neural network simulation environments for the neurosciences. Submissions are welcome through 9/25. (310) 825-2381. [connectionists, 6/12.]

Bruce D'Ambrosio ( is compiling an annotated bibliography of applications of modern uncertainty representation and inference methods in AI, for distribution at a UAI-92 tutorial. BibTex format is preferred, but not required. [Uncertainty List, 6/23.] Here's a chance to mention your publications, with a few short sentences about application area, problems addresses, methods used, and capabilities or knowledge gained. Bruce might even give you a copy of the bibliography.

Aspirin/MIGRAINES developers are seeking short (e.g., two- sentence) descriptions of applications for an upcoming book on "Neural Networks Simulation Environments." They would also like article citations and self-explanatory PostScript figures. Russell Leighton ( [connectionists, 6/23.]

Superstar Teachers Tapes is a good idea whose time might have come. The company pays the most popular teachers at major universities to record their presentations before a live audience in Washington, DC. They have 43 videotaped lectures on psychology, philosophy, religion, and Western culture, with outlines and reading lists for school use. The company plans to record a full college curriculum, plus high school course in algebra, chemistry, etc. (800) TEACH-1-2. [VPN Reports, Channel 38, 6/26.] (Why hasn't PBS made such tapes for public broadcast? Joseph Campbell's series on myth worked well, and I've enjoyed hearing Tom Peters speak. Surely the professors would be willing. NSF might even pay for a science-related series.)

An important copyright decision by New York's Second Circuit Court of Appeals -- the most authoritative circuit for copyright issues -- has established that clean-room procedures can be an adequate defense against software copyright infringement. The function of a piece of code cannot be copyrighted. Altai Software rewrote parts of its Oscar 3.5 scheduler program that had been improperly derived from Computer Associates' source code. The new code performs the same function, but has been judged free of infringement. Computer Associates plans to appeal. [Laurie Flynn, SJM, 6/28.]

Data Addressable Memory (DAM) is a new content-addressable architecture using three RAM memories: 1) A sequential log-book (or "log") memory stores data vectors in the order learned. 2) A "twist" memory interchanges the rows and columns so that a data vector retrieves the most recent log-book address. 3) A "linked-list" memory stores LIFO chains of log addresses that all contain the same data pattern. You query the twist memory for a pattern's address and -- if you need all instances -- chain through the linked list until you reach a zero. Check the log memory if you need to verify that each address does store the data pattern. (The log memory might be unnecessary for pattern-recognition applications.) [Loring Wirbel, EE Times, 5/11.] You can implement this with current software or hardware technologies, but you may have to license the patent rights from Lawrence Dillard of TexTek (Boulder, CO).

LawDesk CD-ROM court rulings are available for New York 2d, California, Connecticut, and Arkansas. Beta testers say that the discs are often faster, more convenient, and cheaper than online information systems. $2,995, or $500 per network plus $50 per user; quarterly updates for $580/year. Thomson Electronic Publishing (Stamford, CT). [Thomas Hoffman, CW, 6/8.]

Philip Leith's new book, "The Computerised Lawyer" (Springer- Verlag, 1991), is reviewed in the 1/92 Computers and Law Magazine. [Pamela N. Gray (, AIL-L, 3/16.]

"The Legal Liability of Information Professionals" is discussed by Martin Felsky in the Canadian Journal of Information Science, #14, 9/89. [John Hughes (, online, 3/27.]

When technologies are first introduced, courts tend to protect practitioners and their new applications. This has happened with expert systems, as in Triangle Underwriters v. Honeywell Inc. (1979) and Chatlos Systems v. National Cash Register Corp. (1979), where courts refused to recognize the tort of computer malpractice. This attitude is changing, as evidenced by malpractice judgments in Diversified Graphics v. Groves (1989) and Data Processing Services Inc. v. L.H. Smith Oil Corp. (1986). AI practitioners would be wise not to claim too much for their systems. [Peter P. Mykytyn, Jr., and Kathleen Mykytyn, AI Expert, 12/91.]

- -- Ken