close this bookVolume 2: No. 28
View the documentNews -- politics
View the documentNews -- economy
View the documentNews -- graphics
View the documentNews -- computer industry
View the documentNews -- OCR; handwriting recognition
View the documentNews -- job opportunities
View the documentDiscussion -- software careers

The San Jose Mercury did a piece on executive compensation in Silicon Valley. John P. Morgridge at Cisco Systems accumulated $80M in stock in just four years. (Increases in the value of previously held stock aren't counted as current compensation, though. There are [several] formulas for computing the future value of current stock options, and these are the basis of compensation reports.) Ten top incomes this year ranged from John Sculley's $6.18M down to Charles Geschke's $1.98M. Stock options are the biggest portion, with cash payments amounting to only $.28M-$1.53M. Executives often get bonuses to sign on, bonuses for performance, and bonuses if they are terminated. Loans to buy houses are often given to cash-strapped hires as golden handcuffs, due if the executive leaves. They may also get loans to exercise their stock options -- after all, the options wouldn't be motivating if they couldn't be exercised. Such loans are often forgiven in stages, much like pension vesting. Silicon Valley executives generally do earn their pay by growing their companies. (Software companies are the least productive, accounting for 2% of Silicon Valley profit and 14% of CEO pay. Networking, telecommunication, and service companies also appear to overpay relative to profits.) Stock is more readily available to employees than it is elsewhere in the country, with discount or revalued stock options used to retain and motivate employees (including executives) during hard times. Critics say that this amounts to giving people a raise because of poor corporate performance. Repricing stock options most helps executives (who own a lot of options), and any new stock issued at the lower price dilutes the equity of investors. [Ron Wolf, SJM, 6/29.]

About 20% of 11,000 employees at Microsoft's Redmond headquarters may now be millionaires, if they held onto their stock options. (Microsoft says it's only about 500 people.) They may be the nation's first large group of millionaires from technical backgrounds. (Most others are executives, lawyers, entrepreneurs, and people with trust funds.) They're not in it for the money; they just want to change the world. They wear blue jeans, avoid country clubs, and may never have bought a house. Their favorite car is a Lexus, the same as Bill Gates. Gates works 16-hour days, and so do many of his employees. Base pay is below the industry norm, but stock-option vesting keeps turnover extremely low. [NYT. SJM, 7/4.]

To judge a company's career environment, watch how employees at your level are treated during a recession. Most companies cut engineering staff when revenues fall. HP instead cut salaries by 10% and cut work time to nine days every two weeks. Employees kept their jobs, benefits, social relationships, and self-esteem. Only a few other companies have done as well. [George Rostky, EE Times, 5/4.]

San Jose's main library is closing its employment-related Silicon Valley Information Center. 22,000 people visited last year, but the library can't afford $80K/year to keep it running. The center was started in 1986 with $895K in city and state grants, and has grown through corporate contributions of annual reports, press clippings, and videotapes. Some materials will remain available, including clips being transferred to CD ROM. [Sherri Eng, SJM, 7/1.] Similar centers could be started elsewhere -- especially if you can copy the database and CD ROMs.

There's an important principle behind Unix, Usenet, shareware, and GenBank: distributed user development. The GenBank genetic database was founded in 1983. By 1987, it had fallen two years behind, and the literature was starting to double every 15 months. Then IntelliGenetics won a $5M/year contract to convert GenBank to a user-maintained system. Free DOS and Mac interfaces were sent to 7,000 scientists, allowing them to enter their own data. (They could also match their data to other entries, so they had reason to enter the sequences.) The database has since grown from 5M base pairs to 78M, and often leads the published literature. [Gary H. Anthes, CW, 6/15.] Lesson 1: If you want to build a service empire, enable users -- or developers, or volunteers -- to contribute most of the work. Lesson 2: Find an angel or cash cow to pay for the coordination effort.

(Apple started with an external development strategy, but shifted most of its support to a few large developers. It lost its vision as "the computer for the rest of us," although the change may have kept the company in business. Microsoft, NeXT, Electronic Arts, and the Free Software Foundation may now be doing a better job of supporting developer communities. Perhaps it works because Bill Gates, Steve Jobs, and Richard Stallman are charismatic; John Sculley is not.)

A SEF Expert System SIG panel recently discussed opportunities for AI developers. Panelists were Bob Fondiller (IntelliCorp), Amos Barzilay (Syntelligence Systems, Inc.), and Leor Jacob (A.I.novation). A report by Dr. William T. Park -- (Gets around, doesn't he?) -- says that nearly all of the remaining expert system tool vendors such as IntelliCorp have moved into C/C++/OOP or CASE mainframe tools. (The C environment at Sun is approaching the power of Interlisp D. Oscar Firschein at DARPA is leading a program to implement C++ on Lisp Machines.) MIS developers are the new expert-system market, except that CBR tools are popular with MBAs. Aion is trying to move into Unix, and it and AI Corp. have been chasing the insurance industry. LISP tools survive mainly at Apple, as the Dylan language. ILOG (from France, pronounced eelog) will be marketing Lisp-based expert systems from a small Palo Alto office. Enfin (San Diego, CA) will be selling expert-system workstation-to-mainframe connectivity tools written in a SmallTalk clone. [SEF, 7/92.] Incidentally, Syntelligence Systems, Inc. is a new company that has purchased the assets of the defunct Syntelligence, developer of Lending Advisor. Major AI labs have downsized or disbanded: ADL, Coopers & Lybrand, Martin Marietta, Price Waterhouse. AI can't compete with cheap labor -- except on speed, uptime, and distributed availability of expertise.

Bill Park also forwarded several net items about crazy working conditions at government agencies. Carl Robinson ( notes that government is conserving budget by hiring contractors rather than civil servants -- even though the GAO warns that long-term costs are higher. Agencies are also starting to cut total workforces through hiring freezes and reductions in contractor workforces. Bruce Salem ( says that an infusion of knowledgeable contractors would benefit the U.S. Geological Survey (Menlo Park, CA). Computing resources, including 50 isolated databases, reflect turf wars between the administrative field geologists. "The only group that has its act together is the group that maintains the seismological data network." The agency produces good printed products, but couldn't provide the same data online. Scott J. Spetter ( tells horror stories from SAIC's Comsystems Division in San Diego. (Other netters have said nice things about SAIC.) Scott calls this group "the most political, cut-throat, abusive of employees, and unethical I have ever seen." He mentions exploitative salaries, refusal to pay referral fees, time-card falsification, swearing, rudeness, hostile treatment, false and damaging personal reviews, lies about layoffs, and one-sentence reference letters (after good references had been promised). [m.j.m, 7/2.] (As middle-management ranks are thinned out, the incompetence of top private-sector executives is becoming more obvious -- and industry is increasingly willing to fire the least competent. That hasn't happened yet in the government and nonprofit sectors.)

Joseph B. Dunphy ( submitted some good advice to He's a mathematician studying stochastic processes, but has found there are few commercial prospects for theoreticians. (He's particularly upset that interviewers have no comprehension of his technical skills. One asked "If you were a flavor of ice cream, what flavor would you be?") If you hope to work in applications, get a second graduate degree in your favorite application area. ("You could point out that we haven't seen one person do EVERYTHING necessary to produce a product since the Paleolithic era. Pointing it out will get you nowhere.") Go to your professor and ask about application areas. Your mathematical background will give you a leg up, so the second degree will be easier than the first. Learn as much computing as you can. Afterwards, you can put out a consultant's shingle to attract [manufacturing] companies who know that they need your skills. (Charge at least $40/hour. Faculty members get $1K/day.) Go to SIAM or other professional meetings for support from professionals in your field. Contact placement firms that advertise in your journals, but don't expect results. Sign up for substitute teaching work ($50/day) to pay the bills. Assemble computers or other consumer goods in your spare time. Avoid personnel departments, placement firms that charge, mass resume mailings, job-hunting books, networking parties, and saturated markets like the financial and service sectors. [Bill Park (, 7/3.]

A job tip from career author Carol Kleiman: "Find out what your boss likes least and take it over." [Newsweek, 6/15.]

-- Ken