|
|
by Peter H. Salus Peter H. Salus is a member of ACM, the Early English Text Society, and the Trollope Society, and is a life member of the American Oriental Society. He has held no regular job in the past lustrum. He owns neither a dog nor a cat. The First Computer Who built the first computer? Zuse? Aiken? Eckert and Mauchly? Stibitz? Wilkes? Atanasoff? Depending on your point of view, any of these men might plausibly be credited. In a previous column, I mentioned Bernard Cohen's biography of Harold Aiken. I've now gone through Scott McCartney's book on Eckert and Mauchly. Cohen writes: "There is general agreement . . . that Mark I heralded the dawn of the computer age" (p. xiii). McCartney says: "John von Neumann didn't invent the computer. . . . The distinction rightly belongs to two men at the University of Pennsylvania, Presper Eckert and John Mauchly" (p. 5). Hmmm. George Stibitz's Bell Labs Complex Number Calculator becomes operational on 8 January 1940. It is demonstrated via a teletype from Dartmouth on 11 September 1940. Konrad Zuse completes his Z3 machine on 5 December 1941—the first fully operational calculating machine with automatic control of its operations. In 1942, with its arithmetic unit of 300 tubes complete, the ABC is abandoned by Atanasoff when he goes into the Navy. In January 1943, the Mark I is operational at IBM's Endicott, NY, labs; it will be moved to Harvard in May 1944. In spring 1945, the ENIAC is "working well." At this point, let me point out that just what a computer is, is less than obvious. The usual criteria are that it be digital, be programmed, be able to perform basic arithmetic functions, and employ a stored program. Many folks reject the creations of Zuse and Aiken because they were electromechanical, not electronic. Eckert and Mauchly's ENIAC was electronic, but it didn't have a stored program. And so we arrive at my conclusion: The winner is Maurice Wilkes; the machine is the EDSAC, designed and built at Cambridge (the one in the UK, not the one in MA). EDSAC ran its first calculation on 6 May 1949. McCartney's fascinating book on Eckert and Mauchly misses this entirely. It seems to flow more from Al Gore's statement (in 1996) that ENIAC was "the first computer in the world" (an indication of the future claim that the vice president had "invented" the Internet) and Bill Clinton's allusion to it in his second inaugural address (January 1997). Read it, but don't believe everything. By the way, I know of no book on George Stibitz nor on Konrad Zuse, though at least Zuse's autobiography was brought out in English by Springer (1993). ARPA's Investments Janet Abbate's history grows out of her 1994 Pennsylvania dissertation. While it is a relatively good read, it also suffers from a mild case of dissertationitis, and under 30 of the references in her 18-page bibliography are from the last five years. Most of what Abbate has to say is correct, except (in my opinion) she goes awry where the OSI/TCP war is concerned. Her academic bias shows on p. 221, where she refers to Hafner and Lyon (1996) as "journalistic" and John Quarterman's (1990) and my (1995) books as not scholarly. My guess is that this means: not published by a university press. Coffee Beans? Coffee Grounds? About two years ago I felt inundated by Java books. The same sort of thing has happened this summer. I received nearly two dozen (some, admittedly, second editions) books on Java, the Java OS, Jini, etc. Too many. But here are my thoughts on three of them. The Java Virtual Machine is an abstract computing machine that enables Java to host applications on other platforms without rewriting or recompiling code. This has been Java's claim for several years. Engel's first-rate book explains how the VM works and how to program it. I especially liked the section on implementing languages (Scheme, Prolog, Eiffel, etc.) and the one on compiling languages. This led me to Sheng Liang's volume on the Native Interface, another excellent exposition on how to integrate code in C or C++ into Java code and how to pass data types. Did Andy Koenig invent the "traps and pitfalls" genre? If he did, I tip my hat to him. Liang's 15-page chapter is very fine. Finally, along comes the Jini spec. Bill Joy put it best: Jini "is designed to bring reliability and simplicity to the construction of networked devices and services." The Jini architecture is relatively simple. The specification is not a simple read, but it is a rewarding one. Telecomms I enjoyed most of Thurwachter a great deal. It's telecommunications without calculus, and it's been a long time since I thought seriously about calculus. The book is clearly intended as a college textbook, but don't let that prejudice you. In our world, if you work with computers, you need to know something about data and telecomms. There's a respectable chapter on ISDN, but it informs me that there are "two main types of switching: Circuit switching . . . [and] . . . Packet switching" (p. 539). Message switching anyone? DSL and ATM are also MIA. The chapter on optics is quite good, but why was WDM relegated to "Multiplexing" rather than "Optical Media"? I have several times bewailed the lack of a good book on optical networks. Never again. Stern and Bala have written a very fine volume on a variety of topics involving optical fiber and transmission. There's a lot of math in this one, but if you give the authors your attention and time, it's well worth your trouble. Finally, there's a second edition of Carne's Telecommunications Primer. Carne covers a lot of ground and does it well. There are nearly 70 pages of acronyms and glossary and a lot of well-presented other information. One area only, OSI/TCP, causes me unease. Carne states, "TCP/IP applications are limited in scope and flexibility; they cannot invoke many of the services that the OSI model envisions" (p. 640). Yep. TCP was designed by folks dealing with real stuff over real networks. They didn't "envision," they made it work. Apology I have received several notes concerning my criticism of Mobility, in general because I said little about the contents but a lot about the presentation (dimensions, typography, etc.). While the book may be ungainly, the contents are first rate, and the editors have done a fine job in pulling the work together. There are a number of papers that are "must reads": Barak and Wheeler, Cabrera, Douglis and Ousterhout, Cheshire and Baker, the late Mark Weiser, Johansen, van Rennesse and Schneider, in addition to the Perkins. Even if (e.g.) IEEE has issued photo-offset 8.5"x11" volumes, there's no good reason for ACM/A-W to do so. Books reviewed in this column:
Scott McCartney
Janet Abbate
Joshua Engel
Sheng Liang
Ken Arnold et al.
Charles N. Thurwachter, Jr.
Thomas E. Stern & Krishna Bala
E. Bryan Carne
Dejan Milojicic et al., eds.
|
 Need help? Use our Contacts page.
Last changed: 23 Nov. 1999 jr |
|
Issue index