Home > General Development > Knuth’s Art of Computer Programming Vol. 4 Knuth’s Art of Computer Programming Vol. 4 Eugenia Loli 2005-02-04 General Development 14 Comments Bookpool.com has posted an excerpt from Knuth’s long awaited The Art of Computer Programming: Volume 4. It is very short and discusses combinatorial searching. About The Author Eugenia Loli Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker. Follow me on Twitter @EugeniaLoli 14 Comments 2005-02-04 9:06 pm We all know that Knuth is god. Do we have to say more? It goes Knuth, to lesser academic MIT/Stanford/Berkley gods to lesser uber phd students, to us mortals that love programming but don’t have pull, to great Microsoft and Sun uber brains. And then we get something. I’m sure i missed something. but this is great. God posts even if we don’t understand.. 2005-02-04 9:09 pm I need to have at least more than one post. God needs to do his compiler book. I wonder if anyone could convince god to release some of his compiler book’ 2005-02-04 9:46 pm It would sure make it more readable for Linux. So much for leveraging his TeX system if his METAFONTS are just going to output Type3 Fonts. Looks fine on Windows but how is this a sign of God if it isn’t Universal? And no Dr. Knuth != GOD. A programming demigod at best. 2005-02-04 11:19 pm Could we link to the actual article and not the slashdot coverage? http://www.bookpool.com/.x/dwtbcyv9d6/ct/163 2005-02-04 11:31 pm I’m not sure why a 14 page excerpt from the book is suddenly newsworthy or even slashdot-worthy… Donald Knuth has been publishing excerpts from the book (or fascicles or pre-fascicles as he seems to call them) form some time. See here: http://www-cs-faculty.stanford.edu/~knuth/news.html for instance. They might be for “beta testing” (as Knuth calls it), but they’re quite interesting already (although I only read the History of combinatorics part yet). 2005-02-05 2:00 am agreed. this is the second article i’ve seen today linked to the slashdot article when a better link could have been chosen. 2005-02-05 3:11 am Of course the lesser gods can take any form they choose. Like Guyus Steelius Junius, who’s currently slumming it on the lowest rank as a Sun employee. 2005-02-05 8:09 am I don’t understand why people call him a ‘god’. For what actually does he deserve to be called a ‘god’ ? His books are of no practical use for the real life software developer/programmer. His books are more or less a collection of math forumlas without any real life explaination. While reading some of his books I realized that this guy is living in his very own world of computers, programming and somehow stuck in the mid 70’s. To much cult and hype around an old guy that somehow made his name become a legend. The same with Richard W. Stevens books, totally overhyped but still declared to be a standard while there are other books half the price written by other nice people which explains more, better and of course covers new topics 2005-02-05 9:05 am Knuth’s books are theory books. They exist to teach you the theory of computer science, not how to quickly solve a specific problem. Since computer science is a branch of mathematics, Knuth’s books are full of math. This is true of any serious computer science work — browse around Citeseer, IEEE, or ACM sometime. Even Microsoft’s research papers are written this way — mathematics is the standard language of computer science. There is a long-standing dichotomy between theorists and practicioners. Computer science is to programming as physical science is to engineering. In both situations, the practicioners apply the theories devised by the scientists. I’m studying for an engineering degree, and I’d say that in a given year, we spend 90% of our time studying theory. Why? Theory allows you to solve new problems — specific examples can only show you how to solve old ones. Who makes their living solving problems that have already been solved? Oh, and Knuth is a god. If only for TeX! 2005-02-05 9:20 am I know what you want to explain I did study myself. Though this still doesn’t make Don Knuth a god or something. There are plenty of good maths books outside and many of them explain the theory far better than the irrational blubbering of Don Knuth. Most software engineering nowadays are build up to ‘solve’ problems ‘quickly’ because of money and the customers who want everything cheap. So there is no time to make indepth background research in mathematical philosophy first. The people who understands Knuth’s books are already experts in maths and in reality don’t really need his books anymore because they are on a high educated level already. These are books written by engineers for engineers or differently said written by experts for experts. 2005-02-05 9:48 am Well, Stevens books are still the best ones written on *practical* Unix and network programming. They contain information and code snippets that can directly be used in applications, something which Knuth’s books don’t. I personally rank Stevens higher than Knuth, but I haven’t read more than small parts of Knuths books. 2005-02-05 10:58 am Knuth happens to study those theories so that others could then “explain the theory far better”. There’s a saying that he invents the algorithms the rest are using. Those books are not for code monkeys, they are for other scientists and trainers of those sad monkeys. 2005-02-05 7:06 pm There is no reason to be derogatory towards the “sad monkeys”. As an engineer, I’m all to aware of the indescribable gulf beteween myself and the physicists and chemists who make my field (aerospace) possible. Yet, I happen to think there is a synergy between my work and theirs — that we both have something to contribute to each others’ work. I don’t think the situation is any different for computer scientists vs “software engineers”. They do different things, but I don’t think one would be nearly as useful without the other. 2005-02-05 7:42 pm > Most software engineering nowadays are build up to ‘solve’ > problems ‘quickly’ because of money and the customers who > want everything cheap. So there is no time to make indepth > background research in mathematical philosophy first. There are enough problems where NOT having the theoretical background will leave you completely clueless. With such problems, you’d not have an algorithm handy that solves the problem within less than ten years. Or you’d have a solution but couldn’t understand it. Of course that won’t happen when you build an application from purchased components. It happens when you build yourself those components that others buy for lots of money. > The people who understands Knuth’s books are already > experts in maths and in reality don’t really need his > books anymore because they are on a high educated level > already. Nobody is “an expert in math” in the sense that he/she knows everything. The more you know, the more you don’t know. Yes, you need some basics to understand the book. But I doubt that many people know everything the book contains. That’s a bit like saying that you don’t need an encyclopedia because you have your common sense and life experience. At some point or the other, you *will* need it or at least wish you had one.