Home Authors Books Subjects Events Software Features Links Newsletter Gifts Blog Write Review What's New

Review - Computers Ltd: What they Really Can't Do - David Harel


Visit bookshop

By tradition in the book trade, negative books don't sell well - so a book telling you what computers can't do should be a turn off. Actually it's got a lot going for it.

As David Harel points out, we are used to thinking of computers as limitless (it's just those irritating programmers that introduce the bugs). He quotes the editor of a computing magazine saying "Put the right kind of software into a computer, and it will do whatever you want it to do... there are no limits to what you can do with software." But the editor was about as wrong as it's possible to be. Even before true computers existed, the mathematical genius Turing had proved that this wasn't the case, and Harel sets out to explain in popular language why we should be aware of real limits on computing.

Harel's computer science is occasionally firmer than his grasp of history. He repeats the often quoted but false origin of the expression "bug" to a large beetle found in the innards of an early computer. It's true that pioneer Grace Hopper did tape a "bug" that messed up a computer into her log - but the term had already been in use by engineers since the 19th century.

The book starts painlessly with an introduction to the nature of computing and algorithms before leading us into the more murky corners of computability. It doesn't quite have the page-turning thrill of a great popular science book, but it remains readable and entertaining most of the time. We find out both about problems that are impossible to solve, and those that are impractical to solve because the time or memory space (or both) to do so are impossibly large. It's remarkable how some apparently simple problems escalate in complexity so quickly that the best computer in existence would take longer than the life of the universe to come to a conclusion.

Even so, there are flaws. Harel doesn't want to impose math etc. on us - which is fine - but he takes it too far, repeatedly mentioning something but not explaining it, presumably because it's considered too difficult. For example, at one point he is talking about a class of problems called linear programming. He mentions an approach to dealing with this called the Simplex method, but doesn't tell us what it involves. Then he tells us that recently there is a potentially better approach - but not what it is. Good popular science avoids confusing the reader with too much technical stuff, but the solution isn't vague overviews, it is good explaining! There are also times when Harel goes on too much about puzzles or spends a long time explaining technical terms that were never really necessary to understand what's going on.

There's a sharply split final decision. The three star rating is for the general reader. If you working in computers or a discipline like operational (operations) research, or you have an interest in meta-maths this is definitely a four star book. This reviewer scores on two out of three of these, and found in enthralling despite having to skip read a couple of parts, and despite being frustrated by the lack of expansion on just what some of these algorithms are about. The general reader will definitely find some aspects of interest, but is less likely to be thrilled.

Also in hardback: Visit bookshop Visit bookshop

Reviewed by Brian Clegg


This site has no connection with Popular Science magazine or other sites and publications with a similar name.

Much of the content of this site is written by popular science writers or friends of popular science writers. Inevitably many of the reviews in such a small community are written by or about someone we know. We always aim to be impartial in our reviews, but there is a connection which we need make clear, as there is no intention to deceive. The content of any review or article is solely the opinion of the author and should not be read or understood on any other basis. The site exists to promote popular science writing and popular science authors and for this reason should be considered promotional material, just as the editorial reviews in an online bookshop or the blurb on the back of a book should be considered promotional.

The website should not be eaten or used where it can come into contact with water.

Disagree with our review? Want to comment on a feature? Contact us at info@ popularscience.co.uk - have your say!

Part of the Popular Science  site

Copyright Creativity Unleashed Limited 2005
Last update 05 June 2007