Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!cs.utexas.edu!sun-barr!decwrl!shelby!agate!saturn!sidney From: sidney@saturn.ucsc.edu (Sidney Markowitz ) Newsgroups: comp.lang.c++ Subject: Re: Book Reports Message-ID: <8095@saturn.ucsc.edu> Date: 23 Jun 89 22:23:19 GMT References: <6590153@hplsla.HP.COM> <6590167@hplsla.HP.COM> Reply-To: sidney@saturn.ucsc.edu (Sidney Markowitz ) Organization: University of California, Santa Cruz Lines: 35 In article <6590167@hplsla.HP.COM> jima@hplsla.HP.COM (Jim Adcock) writes: >Hu, David: "C/C++ for Expert Systems," Management Information Source, >ISBN 0-943518-86-5. 565 pages. $25 > >I spent an hour or two scanning [not reading] this book, and it seems a >strange hodge-podge of Lisp, C, C++, Smalltalk, Prolog discussions, etc. > >I wouldn't recommend this book in general, but it may be useful if you want Here's my review: I found it one of the worst books I have ever come across. My background is Lisp and Expert Systems, including writing the inference engine for a commercial expert system shell (in Lisp). What Hu did in this book was write about his company's (now defunct, I think) Lisp-based expert system shell/toolkit, making an incredibly incomplete and sloppy effort at translating the Lisp into C, and almost ignoring C++. He glosses over all the difficult issues in a Lisp to C translation, like memory management and the symbol data type. He says nothing about real inference engine design, implying that the naive designs you find in primers are how they are done, when in fact they are impractical for real problems due to combinatorially explosive algorithms. There's some code for a parser that can read and tokenize a list of symbols in parentheses, and somehow that gives Hu license to forget that this is a book on C/C++, so by the end of the book there's page after page of examples in Lisp with no justification or explanation of what this has to do with C or how to convert it. In short, "C/C++ for Expert Systems" is not about C, is definitely not about C++, is naive and confusing about Expert Systems, and if it is meant as marketing hype for Hu's company's product, it didn't work. If I didn't feel so strongly about a publisher trying to exploit the hot topics of C++ and Expert Systems I wouldn't even waste my time reviewing the book. -- sidney markowitz