0 of 3 people found the following review helpful:
Interesting ideas endure, December 8, 2001
Mirror Worlds Gelertner 3 starsThe book, first published in 1991 by Oxford University Press, must be read in the context of its day to be fully appreciated. At that time, in the pre-web world, there was a great deal of discussion devoted to Artificial Intelligence (AI) and the Fifth Generation Project driven by the Japanese. If Gelertner had limited his offering to only those topics this book could be left in the pile of such books from that era without loss. Luckily, Gelertner gave us more. While there is much of the book relegated to the AI ideas of that time, there are also insightful and practical observations that have a more lasting appeal. For example, Gelertner delves into the question "What is a program? What does 'software' mean?" Such questions are explored in some detail and other observations are made in the discussions. "Managing complexity must be your goal... we can call it the pursuit of 'topsite'. Topsite-- the understanding of the big picture--is the essential goal of every software builder. It's also the most precious intellectual commodity known to man." We've all heard talk about someone who "sees the big picture." That, according to Gelertner, is "topsight": having perspective, clarity, and a sense of proportion. Why is this important? If we want to have machines (programs) help us see and understand our world (in a "Mirror" of our world), we'll need to teach these machines how to make sense of the information. Minimally, they'll need to be able to sift through the volumes of data and find that data which is "interesting." The very best programs will be able to find those interesting things and present them in a compelling way. All of this demands "topsight." To drive this ideal, Gelertner and his colleagues created "Linda" which serves as the basis for the machinery of such a Mirror World system. The idea is simple: create a Space where information (called a Tuple) can be put, taken, or simply read or examined. Many programs put information in the space. Other programs notice items in the Space, take them, and perform some processing, and put a different item back into the space in its stead. This part of the book, the very practical nuts-and-bolts part, is alive and well today and in active use. While Gelertner's system Linda may not have achieved widespread acceptance, the same idea in another form is quietly thriving: JavaSpaces. The same notions described by Gelertner to support his Mirror World now serves as the heart of many commercial applications. Gelertner has a lot to say. Yes, some of it now appears dated and some of the ideas he touts have been discredited. But, nobody said predicting the future was easy business! My recommendation is thus: forgive Gelertner the detours he takes (that we all took) and find within the book all those things which have inspired--and will continue to inspire. There are ample enough thoughts within those pages to make the time invested in a careful reading well worthwhile.
|