|Volume 1: No. 03|
Sun Microsystems (Mountain View) has announced formation of Sun Labs, a long-range research group to be headed by Wayne Rosing. Sun is also forming two marketing subsidiaries: Suntech Enterprises to sell Sun software and Sunsoft, Inc. (under Edward J. Zander) to be the "Microsoft" of the UNIX world. [Jon Kennedy in High Technology Careers, 4/91.]
SuperMac (Sunnyvale) is spinning off a software division with Laurie Girand as head of software development. [Wendy Woods, Newsbytes/Bay Area Computer Currents, 3/26/91.]
Pat Hayes, AAAI president-elect, has moved from Xerox PARC to the MCC Cyc project under Doug Lenat. The current Cyc effort, funded at $3.5M/year, includes 25 inference engines for drawing conclusions from a massive knowledge base. [Harvey P. Newquist III, AI Expert, 4/91.]
Futurist John Naisbitt forecasts a decade of decentralization, small companies, and telecommuting. Naisbitt's own home is in a 9,000-foot Colorado village of 1,400. [Ian Stokell, Newsbytes.]
A careful study by the National Institute for Occupational Safety and Health has exonerated video display terminals (VDTs) of any effect on pregnancies among 2,400 telephone operators. The study did not cover other suspected health risks.
LIBRES (Library and Information Science Research Electronic Conference) is a new weekly discussion list covering library and information science research. LIBRES will include discussions of research in progress, reviews of research, queries and responses from participants, and conference announcements. Biographies of list members will be available on the LIBRES fileserver and via anonymous FTP from ksuvxa.kent.edu. You may subscribe by sending a "SUB LIBRES YourFirstname YourLastname" message to email@example.com or firstname.lastname@example.org. Send questions or submissions to email@example.com. The co-editors are Diane Kovacs, Tom Froehlich, Julie McDaniel, Amey Park, Barbara Schloman, and Rosemary Dumont of Kent State and Ellen Detlefsen of UPittsburgh.
The Science and Technology Information System (STIS) at the National Science Foundation is a free, 24-hour electronic information dissemination system for NSF publications. The full text of publications can be searched and copied. Publications include the NSF Bulletin; Guide to Programs; grants booklet (GRISE); program announcements; press releases; NSF telephone book; reports of the National Science Board; abstracts of funded research projects [not always accurate!]; and international news. Contact STIS, NSF Office of Information Systems, Room 401, 1800 G Street NW, Washington, DC 20550; firstname.lastname@example.org; (202) 357-7555, Fax (202) 357-7745. (On the Internet, access STIS via "telnet stis.nsf.gov" with login name "public". For dial-in access at 1200, 2400, or 9600 baud, call (202) 357-0359 or (202) 357-0360 with even parity, 7 data bits, full duplex, and VT-100 emulation. Type "[ENTER] . [ENTER]", then login to STIS. The interface will give you a taste of what it's like to work at NSF. Use your arrow keys plus ESCAPE, TAB, PGUP (^U), PGDOWN (^D), TOP (^T) and BOTTOM (^B). You may have to use RETURN or "+" keys instead of ENTER, or a numeric keypad instead of arrow keys.)
The National Ad Search ((800) 457-8482) publishes weekly reprints of ads from 75 regional newspapers. Their newest service, Ad-Fax, will send you a fax of one job category for $15, or $50 per month. The fax reaches you several days before the newspaper would, but still a few days after local publication. [Jim Lynch, Knight-Ridder.]
Contact Tony McEnery (email@example.com) for information about the new Journal of Applied Computer Translation. ACT Vol 1.1 includes a call for commercial software descriptions. [Sigma Press, 1 South Oak Lane, Wilmslow, Cheshire, SK9 6AR, UK.]
The NL-KR discussion list (firstname.lastname@example.org) has announced a new ACL-supported Consortium for Lexical Research under the direction of Yorick Wilks, with guidance from Roy Byrd, Ralph Grishman, Mark Liberman, and Don Walker. The Consortium is an organization for sharing lexical data and tools used for precompetitive research on natural language dictionaries and lexicons, and for communicating the results of that research. Members will contribute and/or use resources in order to perform their research. CLR will negotiate agreements with providers. Major funders of work in this area in the U.S. have indicated interest in making participation in the CLR a condition for financial support of research. An annual fee will be charged for membership, with the Consortium soon becoming self-supporting.
CLR solicits contributions of all kinds. Data: word lists (proper nouns, count/mass nouns, causative verbs, movement verbs, predicative adjectives, etc.); published dictionaries; specialized terminology, technical glossaries, etc.; statistical data; synonyms, antonyms, hypernyms, pertainyms; phrase lists; etc.. Tools: lexical database management tools; lexical query languages; text analysis tools (concordance, KWIC, statistical analysis, collocation analysis, etc.); SGML tools (particularly tuned to dictionary encoding); parsers; morphological analyzers; user interfaces to dictionaries; lexical workbenches; and dictionary definition sense taggers.
As part of repository management, CLR will provide a catalog of, and act as a clearinghouse for, utility programs; compile a list of known mistakes in each of the major published sources; set up a new memorandum series; act as a clearinghouse for preprints and hard-to-find reprints on machine-readable dictionaries; conduct workshops, including an inaugural workshop in late 1991 or early 1992; and provide a catalog for access to repositories of corpus-manipulation tools held elsewhere. Contact The Consortium for Lexical Research, Rio Grande Research Corridor, Computing Research Laboratory, New Mexico State University, Box 30001, Las Cruces, NM 88003; email@example.com (or firstname.lastname@example.org); (505) 646-5466, Fax: (505) 646-6218.
The Natural Language Software Registry is a catalog of research and commercial NLP software, including:
- speech signal processors such as the Computerized Speech Lab (Kay Electronics) - morphological analyzers such as PC-KIMMO (Summer Institute for Linguistics) - parsers, such as Alveytools (University of Edinburgh) - knowledge representation systems such as Rhet (URochester) - multicomponent systems such as ELU (ISSCO), PENMAN (ISI), Pundit (UNISYS), SNePS (SUNY Buffalo), - applications programs (misc.)
Contact Director Elizabeth Hinkelman (email@example.com) for information or reviews. Send for the questionnaire form if you have NLP software, documentation, or reports to contribute. NL Software Registry, Center for Information and Language Studies, 1100 East 57th Street, Chicago, IL 60637, USA. [From NL-KR.]
The Nijmegen Corpus consists of 130,000 words of modern British English, with a full syntactic analysis of each utterance. Contact Hans van Haltern (firstname.lastname@example.org) of the TOSCA Group, Dept. of English, U. Nijmegen, for details about the Corpus or about Nijmegen's language-independent Linguistic DataBase (LDB) software for accessing such resources.
The following were among the tools and prototypes scheduled for demo at RIAO 91 ("Intelligent Text and Image Processing") in Barcelona early this month [Ed Fox (email@example.com), NL-KR]:
- ALCESTE text-analysis system (IMAGE, France) - PRIAM real-time news bulletin analysis (SY-MEDIA, France) - AMI French/German message sorting (Thomson RCC and CORA, Fr) - STO3/STO4 Speech Recognizer System (G. Bekery Acoustical Research Laboratory, Hungary) - SGML/Search retrieval system (Berger Levrault--Advanced Information System, France) - ZEN X-Windows hypertext system (BULL, France) - GUIDE, IDEX (OWL, Great Britain) - TOPIC graph-based document retrieval system (Verity, USA) - MULTIMEDIA INFORMATION SYSTEM (IN TECS, Italy) - TAKE 5 multilingual pseudo-NL retrieval system (EDIAT, Fr) - SPIRIT information retrieval system (SYSTEX, France) - DOXIS structured/full text system (ERGOSUM, France) - STATUS/E structured/full text system (HARWELL COMPUTER POWER LTD, Great Britain) - PSIDOC software tools for documentary databases (JOUVE, Fr) - MOVIE retrieval with interactive decision making (COSM, Fr) - ILIADA retrieval and library management with full-text and hypertext support (Software AG, Spain) - MULTIMEDIA SYSTEM for teaching biological sciences (CTU Univ. of Milan, Italy) - METAL machine-translation system for several languages (Siemens, Germany, Spain) - TEXIRIS 2 PLUS Omnifont OCR with 12-language dictionary (Image Recognition Integrated Systems IRIS, Belgium) - READSTAR OCR on a TRANSPUTER card (Inovatic, France) - TEXTPERT OCR for MacIntosh and Windows (CTA, Spain) - Intelligent Information access to 200 data banks in Europe and the U.S. (INFOTAP, Luxemburg) - EPOQUE access to European Patent Office Databases (European Patent Office, CEE Holland) - IMAGEDB image/descriptor database management (CSI, Spain) - AUTOMABB PC multimedia authoring system (MABB Sistemas Interactivos, Spain) - INFODOC document storage, image processing, and retrieval (INFODOC SA, Spain) - CLARITY text/image archiving system (Micronet, Spain) - HYPARCHIV Windows hypertext (ACS Systembereitung, Germany)
Review -- SKE knowledge-engineering methodology.
Karen N. Gardener of the Bechtel AI Institute gave a talk on Structured Knowledge Engineering at a recent AIA meeting. SKE is a software-development methodology with enough structure and documentation to be acceptable to MIS managers. Bechtel AI Institute is teaching short courses based on the work at BOLESIAN, B.V. in The Netherlands, a firm that will soon grow from 35 to 300 knowledge engineers. [AIA is the Artificial Intelligence Association, a monthly seminar group that meets in Palo Alto. The contact is Leor Jacob, (408) 734-5760. Bechtel is at 50 Beale St., P.O. Box 193965, San Francisco, CA 94119-3965.] SKE diagrams are similar to Entity-Relationship diagrams, but with "inference boxes" and information-flow arrows replacing relationships. These diagrams represent the implementation of "tasks" in terms of "inference modules," two of the four knowledge layers that are iteratively refined during algorithm design. A top layer addresses control issues; a bottom layer maps between SKE terminology and hierarchical domain-knowledge representations. Once the conceptual design is complete, implementation can take place in any computer language or shell. (Bechtel is currently evaluating the strengths and weaknesses of different expert-system tools for implementing each type of knowledge representation and inference process.) A strength of this approach is the catalog of two dozen or so standard inference modules (categorization, diagnosis, selection, refinement, etc.) Anyone learning to think in these terms would indeed be helped in high-level algorithm design. It's a sure cure for programmer's block. A weakness is that the repertoire has not yet expanded into continuous domains such as signal processing. All of the examples I saw were suitable for discrete manipulations (financial applications, object-oriented programming, logical inferences). That's fine for the bulk of all programming, but I saw nothing that would support the design of neural networks, blackboards, or other signal-processing techniques. Karen also seemed to find clustering and multidimensional mapping techniques difficult to think about in this framework, except as black-box "heuristic matches." Low-level knowledge relationships were quite restricted, consisting only of attribute, sequence, is-a, and caused-by. Representing different types of is-a relationships could be quite difficult. (Clyde is an elephant. Elephant is a species. Therefore Clyde is a species.) My conclusion is that SKE will indeed be an important methodology, offering a bridge between MIS and AI people. I'm no authority on systems analysis and functional decomposition, but SKE appears to be an attractive alternative. It is certainly preferable to software design approaches that are limited to a particular expert-system shell. I do regret, though, that Karen was so negative about AI's rapid-prototyping approach. Prototypes are very useful exploration and marketing tools, and I would hate to see them banned in software development contracts.
Daniel D. Corkill (firstname.lastname@example.org) is President and founder of Blackboard Technology Group in Amherst, MA. He earned his Ph.D. in Computer and Information Science from UMass in 1983, having done pioneering research in cooperative, distributed problem solving. Dan rose to Senior Research Computer Scientist at UMass, having headed two major AI software development projects: UMass Parallel Common Lisp and the GBB blackboard system. (His company now sells and supports GBB 2.0, a toolkit for blackboard-based applications. GBB is an extension of Common Lisp and CLOS, inheriting the environment and tools of the host implementation.) Dan's research interests continue to be in these areas, including planning and control in large AI systems and in design and programming methodologies for AI system building. He has presented tutorials at AAAI and IEEE CAIA, and was chairman of the AAAI 3rd and 4th Annual Workshops on Blackboard Systems. His company helps clients design, develop, and install blackboard systems. They also teach short courses -- with Dan in the lead, of course. Call (413) 256-8990 to get on the mailing list. (The next course is June 10 -13, with a $200 discount for payment by May 20. If oversold, there may be another course the week before or after. There's a first-day-only option for just $395, including accommodations.)
Colston Sanger (email@example.com) is willing to share his experience with GID, a software engineering, UNIX and open systems consulting firm nominally headed by Professor Manny Lehman of Imperial College, London. (Manny is retired now.) GID is also extending into GIS and visualization, document image processing, groupware HCI, and network management. Colston and others are essentially still a startup: self-funded, keeping overheads to an absolute minimum. He strongly recommends forming a confederation with others to share costs and help each other out.
Working from home is not a problem, but you really need good connectivity: e-mail, network news, phone, fax, answering machine. Also a proper office chair with back support, which is much more important than a fancy desk. [See the April Spectrum for suggested workstation layout. If the chair has arms, make sure your elbows don't rub when you type. A hot water bottle makes a great footstool in Winter, and with a lap robe you hardly need to heat the house. -- KIL]
At an office, letters just get posted, copies just get done. As an independent, queuing behind all the old codgers at the post office is really irksome. Going out to the copyshop and finding it closed for lunch is also exasperating. But Colston enjoys the exercise.
Marketing is a problem. So is image. You have to look bigger than you are, and feel as if you mean to be around for the long haul. You can't get away with just b/w laser print on Conquerer stationery. GID invested in graphic design by a friend. [After a friend gives you a discount, can you reject the design? Beware especially of low-cost legal advice from relatives and their friends; it may be less than a best effort, or an ego trip outside the giver's area of expertise. -- KIL]
- A4 card folder, embossed logo, solid green inside. Can be used as a wallet for the company profile or, split down the crease and then bound, as a front and back cover for proposals and reports; - writing paper with green logo and printed addresses; - the same, but just with logo - a general purpose sheet; - plain continuation sheet; - compliment slip; - envelope with logo; - business cards.
The general-purpose sheet gets used for the pages of the company profile, as a fax header, for invoices, forms, etc. Everything presents a consistent image.
Indirect marketing is long-range, name-awareness stuff. GID is surprisingly well-known in the UK for such a small company. That's because they have taken pains to join and be involved with things. Getting in as subcontractor to a lead supplier for major account bids also helps. Building relationships takes time, but you have to do it. It also helps to speak at conferences and to publish papers and books. (Colston has a book just out on Open Systems and the single European market after 1992; Chapman & Hall, ISBN 0 442 31327 6.)
GID doesn't do much direct marketing -- the cold letter or phone call. When they do, either they don't follow up at all (just aiming for name awareness) or they send one or two letters in a week and follow up intensively. Most of their business comes by recommendation, by personal contact (old friends and colleagues who have moved on, recommendations from clients, repeat business) or from the indirect marketing. [Colston Sanger; GID Ltd, 69 Kings Road, Haslemere, Surrey GU27 2QG, UK; Tel/Fax: 0428 654821.]
Commerce is no shame. You engage in commerce every time you accept a paycheck for your services. (Both parties benefit. You can expect your employment to end if either side sees no advantage in continued commerce.)
When I was in college, I thought of business as a purgatory for failed academics. Successful students went on to graduate school, then became tenured professors. Now I see that professors also differ in status, with the project leaders, department heads, and institute directors generally being those who know how to sell. Industry likewise values and rewards those who can market ideas or services to bring in revenue.
I'm no entrepreneur. Empire building -- with little concern for the product sold -- just doesn't turn me on. Mass-market sales figures are not my measure of success. Fortunately, I've learned that business principles also apply in broader contexts.
As an NSF program director, I saw plenty of sales pitches. Most were incompetent by business standards. Professors often state what they want to do without saying why the Government or the taxpayers should care. A good, business-like proposal should state what you want to do, why we need it done, and why we should believe that you can do it.
If you're supporting yourself in the marketplace, you need a SERVICE (or product) that you can offer a CLIENT in order to solve his PROBLEM. The three must come together. Academics often focus on abstract problems, and hence profit minimally from their own efforts.
It's difficult to become expert at solving a class of problems, then find a customer. The world is so complex that you will never reach a sufficient state of readiness -- and, if you did, you would find that other consultants and engineers were already serving your customer base.
If you can't market yourself as a product, what is your value? Time. You have time to spend, and can spend it efficiently, on problems that your client doesn't know how to address. You don't need to have all the answers up front. (If you have the answers, in fact, you won't stay employed for very long.) Whether you work on the client's problem as a consultant or as an employee is really just a matter of mutual preference.
Finding someone with a problem is easy, but finding a paying client with a problem you care to work on can be hard. Vertical markets (e.g., finance, insurance, printing, blowing people to bits) are the easiest to serve. One way to get started is to talk with many businessmen in one domain, then sell the techniques you've learned to the least-advanced subset. (This is a good way to become a computer systems consultant.)
If you'd rather serve a horizontal market, try the following. State your research interest as "I've always wanted to be an X" or "I've always wanted to do X." Now change it to "I want to serve X." You have the interest and the background. Why not? This newsletter serves computer and information scientists. Every industry has newsletters, associations, conference organizers, industry analysts, reporters, authors, editors, publishers, etc. Develop activities you enjoy into a job where hard work and good service relate directly to increased income.
You will need a written business plan, funding, business training -- all the usual hurdles. Choosing a price/volume tradeoff is the first decision: will you develop a product at your own risk and then sell to many customers, or offer a service and sell each hour only once? (Newsletters and short courses are somewhere in the middle.) Know your competition, of course, and respect Murphy's Law. Consider an apprenticeship or intervening job to gain needed skills.
Even if you choose corporate employment, you should clearly understand why your boss or client is paying you. Get to know your customer's customers, because it is his (or her) sales to them that pay your own salary. Once you understand the relationship, make sure your boss also understands it. And don't sign on with anyone who isn't 100% dedicated to serving the needs of the paying customers.