BKA-LIFE.RVW 931221 Random House/Vintage 201 E. 50th St., 31st Floor New York, NY 10022 212-751-2600 800-733-3000 800-726-0600 or 400 Hahn Road Westminster, MD 21157 "Artificial Life", Levy, 1992, 0-679-73489-8, U$13.00/C$17.00 One of the strengths of the LOGO computer language is its suitability for program constructs involving recursion. Many programming languages have a number of forms of iteration whereby a routine in the program may loop a number of times. In recursion, however, separate copies of the program start up using as input the output of the parent process. The program, in a sense, is alive, and reproduces itself. Using this tool it is remarkably easy to generate both graphical and text objects which ape the products of living organisms with startling fidelity. It is fascinating to do and watch ... but ultimately rather pointless. The same could be said of this book. Levy has collected anecdotes from the various areas, mostly computer related, where researchers are trying to imitate various aspects of life. The material is entertainingly put together and easy to read. If you wish an introduction to a subset of the people working in these fields, this is a quick overview that might do it. If, on the other hand, you are interested in the technology itself, forget it. Other than some outcomes of experiments, there is almost no technical information at all. No analysis, either. In the discussion of genetic algorithms, where programs mate, mutate and are then tested for fitness determined by a certain task, Levy makes much of the fact that these programs essentially write themselves. This is, however, simply an extension of the old Monte Carlo algorithm, whereby the testing of random data eventually gives you an approximation of the answer to a problem. Both the Monte Carlo and genetic algorithms are severely restricted in the problems they can be used to solve. I have been too harsh in my dismissal of the amount of analysis in the book. Levy does raise the questions that need answers. Such as: What *is* life? How can we test these systems? Can we control artificial life if we do create it? Might artificial life be too "expensive" in terms of resources and energy? One vital question, though, is never considered. When one examines an apparently life-like machine, one must then decide if a machine can truly live, or only ape life. If one asserts that machines can live, one is then beset with a related question: Are we merely biological machines and nothing more? Levy appears to assume the answer is yes, without ever having raised the question. In which case his condemnation of programmers who look forward to our replacement by living machines is decidedly odd. My primary interest, of course, is in the last chapter, "The Strong Claim." The strong claim for having created artificial life is, as might be expected, the computer virus. Unfortunately, the strong claim is not supported by strong facts. While the section on Fred Cohen's research is interesting, Levy has done little additional research. His history is quite careless with facts claiming that the "cookie" prank was a virus. Core Wars is also claimed to be viral and the 1984 Scientific American article is said to have influenced programs known to have been written years before. The author of Den Zuk is said to be Venezuelan, rather than Indonesian, and the derivation of the name is wrong. The discussion of viral programs is limited and seems to be merely a lead-in to talk about the morality of artificial life. This latter subject comprises half the chapter and, except fleetingly, makes no reference to viral programs at all. This, and the earlier discussion of Von Neumann machines which reproduce themselves using materials at hand, bring to mind a Greg Bear novel which is based upon the destruction of the earth by self-reproducing machines. The controlling galactic government considers the construction of such machines automatic grounds for the extermination of a species. Many in the virus research community might agree, noting that it is inherently impossible to know what damage self-constructing automata might cause in an unknown environment. Levy touches on this tangentially in places, but never really examines it. The book is an amusing diversion. One cannot say that it overstates its claims because, in the end, it makes no claims. In the end, we are left with the questions, some interesting biographical points and little else. To quote Marvin, "Life. Don't talk to me about life." copyright Robert M. Slade, 1993 BKA-LIFE.RVW 931221 ====================== DECUS Canada Communications, Desktop, Education and Security group newsletters Editor and/or reviewer ROBERTS@decus.ca, RSlade@sfu.ca, Rob Slade at 1:153/733 Author "Robert Slade's Guide to Computer Viruses" (Oct. '94) Springer-Verlag