Dr. Dobb's Journal May 2000
The future of reading is the subject of an exhibition at the Tech Museum of Innovation in San Jose, California, running through September 7, 2000. Sponsored and developed by Xerox Palo Alto Research Center (PARC), "Experiments in the Future of Reading" exhibits a number of projects that combine reading with the processing power and display technology that is becoming available for multimedia applications.
The "Reading Eye Dog" (RED), for instance, reads documents and immediately converts them to speech. RED is a metal dog with two camera lenses for eyes and an embedded computer optimized for text-to-speech processing. Put any page of text on a stand in front of RED and, in a matter of seconds, he reads the text out loud. RED could be a powerful application for the blind.
Likewise, the "Listen Reader" provides interactive soundtracks to go along with children's stories. The Listen Reader is nothing more than an easy chair with built-in speakers and a reading tablet that swivels above your lap when sitting in the chair. The tablet has pages like a book, but when you move your hand over illustrated areas of the page, different sounds corresponding to the content of the page are activated (the rush of a waterfall, the barking of a dog, and so on).
Other exhibits include a walk-in room-sized digital comic strip and technology for embedding digital information into ordinary printed pictures. For more information, see http://www.thetech.org/ and http://www.parc.xerox.com/xfr/.
A Harvard Medical School/Massachusetts General Hospital study indicates that about 12 percent of the over 2000 scientists surveyed at a number of medical schools have been denied access to research data produced by their colleagues -- particularly if the research involves a patent application (see http://www.mgh.harvard.edu/DEPTS/ pubaffairs/Releases/021600secrecyinscience.htm). Consequently, concern over how the cross-over between academic research and commercial implementation is affecting the open exchange of knowledge in academia continues to grow.
With the promise of big bucks from venture capital, it is no surprise that data withholding is becoming more and more common. For example, Washington University in St. Louis recently issued a press release announcing that Cisco Systems had acquired a company started by three university computer scientists for a price of $355 million. Two of the scientists are still active professors and are also executives in the acquired company. DDJ recently attended a seminar at Stanford University, at which the presenter was both a university professor and chief technologist for a Silicon Valley startup. When a student asked a question about the inner workings of the technology the professor was demonstrating, the "professor" declined to answer, explaining that the "chief technologist" didn't want to give away the company's trade secrets.
When confronted with this issue, university administrators tend to shrug their shoulders, since universities also benefit from royalties that technology-transfer programs generate. But if the mission of higher education is to disseminate knowledge, you have to wonder where protecting trade secrets stops and conflict of interest starts.
Researchers at the Oak Ridge National Laboratory (ORNL) think optical transistors may one day replace electronic transistors. In a recent prepared statement, Panos Datzkos of ORNL's Engineering Technology Division, states that "what we have is a novel approach for optical switching that might optimize all four parameters simultaneously; efficient, low-cost, fast, and large-scale integration of an all-optical switch will revolutionize computing much as fiber optics have revolutionized the telecommunications industry." The ORNL project uses a diode laser LED that causes optical absorption in the waveguide material. The strain created by the absorption of light causes the waveguide to move. This motion can be used as an on-off switch, redirecting light from one end into two or more channels, which is the equivalent of a switch or modulator used in conventional transistors. See http://www.ornl.gov/news/ for more information.
If the Y2K problem was caused by the shortsightedness of computer programmers, there will be no such excuse when the calendar approaches the year 5000. That's because David Book, a physicist at the University of Maryland, is giving us ample warning about a potential mathematical discrepancy in the Gregorian calendar. The modern calendar is supposed to match the time between two vernal equinoxes. The vernal equinox marks the start of spring and occurs on the day when the Sun rises over the Earth's equator (thus dividing night and day into almost equal length). Using this definition, a year is 365.2422 days. By adding leap years at regular intervals, the average year turns out to be 365.2425 days, which is close to the time interval between vernal equinoxes, but off by 0.0003 days.
If nothing is done to correct this discrepancy, our calendar will be a day ahead of schedule in 2915 years. Book calls this the "Y5K problem" since it will occur near the year 5000. There have been some proposals to fix this problem. One is to eliminate leap days in the leap years divisible by 4000 (such as 4000ad and 8000ad). But then again (as some programmers probably thought about the year 2000 back in the 1960s when coding two-digit dates in Fortran), it's not our problem.
The Association for Computing Machinery (ACM) has granted its 1999 ACM Allen Newell Award to Nancy G. Leveson for her pioneering work in creating the discipline of software safety. According to an ACM prepared statement, Leveson's book Safeware: System Safety and Computers (Addison-Wesley, 1995, ISBN 0201119722) "is used by practicing engineers in many engineering disciplines and in classes not only in computer science, but also in industrial, mechanical, and other engineering departments. Her contributions within computer science have also been broad, and have spanned the areas of fault tolerance, software engineering, human-computer interaction, and formal methods."
The formal modeling and analysis techniques introduced by Leveson are used world-wide for quality assurance of software for nuclear power, commercial and military aircraft, medical devices, and various transportation systems. Leveson received her Ph.D. in 1980 from UCLA and spent a large part of her teaching career at the University of California, Irvine, before teaching at the University of Washington and now at M.I.T. Her home page is http://sunnyday.mit.edu/.