From: utzoo!decvax!harpo!npois!ucbvax!C70:human-nets Newsgroups: fa.human-nets Title: HUMAN-NETS Digest V5 #67 Article-I.D.: ucb.1110 Posted: Tue May 18 10:41:56 1982 Received: Thu May 20 02:14:07 1982 Reply-To: s >From G.MDP@Utexas-20 Tue May 18 10:28:42 1982 HUMAN-NETS Digest Tuesday, 18 May 1982 Volume 5 : Issue 67 Today's Topics: Query - Nomic Players & Braniff Intrigue, Programming - Dijkstra & Languages for Good Programming ---------------------------------------------------------------------- Date: 16 May 1982 1556-CDT From: David PhillipsSubject: Nomic, "The Paradox of Self-Amendment" game I am intrigued by Nomic, the game by Peter Suber described in the June 1982 METAMAGICAL THEMAS column in ``Scientific American''. I would enjoy a chance to play it with others. I've entered the ``INITIAL SET OF RULES'' in file: Nomic.Doc on UTEXAS-20. You can FTP the file by logging in as ANONYMOUS. ------------------------------ Date: 17 May 1982 1705-PDT From: Craig W. Reynolds from III via Rand Subject: AA hacked BRANIFF? Does anyone out there know anything about the charges by the Braniff exec that unnamed persons at American Airlines had hacked Braniff's flight reservation computer system? They also alleged other nasty business practices (such as "jaw boning" Braniff's bankers). Specifically it was stated that data on the computer system was modified to indicate that scheduled flights did not really exist, and deleted some passenger reservations. -c ------------------------------ Date: 17 May 82 16:41-PDT From: rubin at SRI-TSC Subject: Dijkstra's Ego Despite the lofty tone of his writings, Professor Dijkstra is anything but egotistical. I believe the Good Professor is really quite a humble and self-effacing man; his writing style simply belies his true nature. I feel we should offer not flames but our forbearance for a problem that Dr. Dijkstra must understand all too well. It is practically impossible to teach good writing to students that have had prior exposure to Dutch: as potential writers they are mentally mutilated beyond hope of regeneration. --Darryl ------------------------------ Date: 18 May 1982 00:58-EDT From: Keith F. Lynch Subject: Dijkstra I didn't see any positive comments about any language from Dijkstra. I wish he would tell us what computer languages, if any, he considers useful, or at least harmless. ...Keith ------------------------------ Date: Monday, 17 May 1982 09:33-EDT From: Jon Webb Subject: Dijkstra I have been taught Structured Programming by people following Dijkstra's approach at Ohio State and the University of Texas at Austin, and also I have listened to some lectures by Dijkstra at those places, and I must say that there is definitely something there. Programming in the way Dijkstra advocates leads to a much deeper understanding of the algorithm, and can often lead to a more elegant or more efficient algorithm. This is especially true when the programs compute number-theoretic or bottom-level operating system functions. The problem with applying Dijkstra's approach to more complex problems, like user interfaces, is not in the methodology but in the ill-defined nature of the problem to be solved, and the fact that the problem must be solved regardless of the elegance of the algorithm. Jon ------------------------------ Date: 17 May 1982 1706-PDT (Monday) From: davidson Subject: good BASICs It's rather tiresome to read about all of these totally incompatible "good" BASICs. Even if portability is not an issue (do you really want to reinvent the wheel constantly?) everyone should know that nice control constructs (WHILE, REPEAT ... UNTIL, etc.) are not what Pascal (and Pascal derived languages) are about, and is very little of what constitutes the discipline of structured programming. ALGOL had those constructs, but Pascal has handily replaced it. The strength of Pascal is in the data structures, and in the compile time type checking. However, I wish to emphasize that structured programming is not dependent on the programming language used. In fact, until the design of a program is nearly complete, it should be in English. Pascal's virtues, then, are two: (1) making coding easier, and (2) making bugs harder. Greg Davidson ------------------------------ Date: 17 May 82 8:27:31-EDT (Mon) From: Dave Farber Subject: Programming Languages Sounds to me that we are mixing up a lot of things. I have always taken the position that a person who claims to be a professional in this field should have a selection of tools. In one part of the computer business that means a selection of languages. I speak Basic, Pl/1, SNOBOL, Fortran, Pascal, Modula, Ada, Lisp etc and have a working acquaintance with several others (even IPL V). The fact that I first learned the 650 L language seems not to have damaged me beyond hope. The main problem in my mind is knowing when to use what langauge. To do string manipulation in Fortran is difficult while to do floating point calculations in SNOBOL is rather foolish. Again there is a need for many tools and people who know when a particular tool is applicable. Dave By the way, a person who knows how only one machine is programmed at machine level is illiterate in this field also. ------------------------------ Date: 17 May 1982 1217-EDT From: PETER MILLER AT METOO Reply-to: "PETER MILLER AT METOO in care of" Subject: Anthropomorphic Languages (Truths That Might Hurt) One of the greatest truths that has been learned from the development of programming languages and programming systems is that there is no single language or programming technique that is perfectly suited for every problem. The greatest problem faced by non-programming professionals in attempting to use computers is the mapping problem - how to state and solve their problem (which is well-understood in their own internal model) in another quite alien model. I would agree with Bruce Lucas that I would rather have a thoroughly rigorous, mathematically-oriented language (probably programmed by mathematicians) for problems such as FFT, string-matching, etc. Other problems - office-oriented information systems, and business data processing - seem less well-suited for such languages and programmers. Anthropomorphic languages, as real production tools, are really in their infancy. Precision is possible. Even building good software engineering practice into such a language is possible. Model-based programming with natural language-style syntax offers the potential of supporting a larger programmer base than is currently possible. Regardless of the elitist contempt that EWD holds for such technologies, they will be given their opportunity to compete in the marketplace. Peter B. Miller ------------------------------ End of HUMAN-NETS Digest ************************