Book: Things A Computer Scientist Rarely Talks About

July 24, 2007 01:37 AM

OK, so I've been meaning to post about this for a while, but never got around to it. I picked up a book called Things A Computer Scientist Rarely Talks About by Donald Knuth a few weeks ago in Borders. It was really more of an impulse buy, which is odd coming from someone who doesn't buy a lot of books, because I know I'll never read them. (I don't read much of the stuff I'd like to read; there's very little chance that I'm going to take time to read stuff I only bought on impulse.) But this one looked really interesting--well, after I actually read the blurb about it on the back.

I'd like to say I had a better reason for picking it up than I did. It would certainly make me look smarter and more cultured if that were the case. But I actually misread the title on the shelf as "Things A Computer Scientist Really Talks About." So I thought it was going to be one of those how-to-think-like-a-programmer books, which is why I initially picked it up. Also, the title was the only thing I noticed. I've heard of Donald Knuth, but haven't read anything he's written and probably wouldn't be able to sum up his contribution to programming any more than "he wrote books about programming."

The book is taken from a series of six lectures Knuth gave about religion and computing. The topic isn't as speculative as it sounds--it's not a book about how new technology is going to impact our lives and ethics and morality, as one would probably assume from hearing the words "computing" and "religion" in the same sentence. Rather, it's a more personal approach. That's not to say there's not a lot of math and logic in the book--there is, and it's very much over my head. But it's the details, not the heart of the work.

In addition, this isn't a preachy book. Knuth is certainly not your modern fundamentalist. And I think that's what makes the book so interesting--the discussion is all about the journey, not any particular conclusion he's trying to support. In that sense, it's very sincere. It's clear he's speaking from a perspective in which faith and reason (and other subjects) can be completely intertwined in the same discussion; one does not have to overshadow the other. (Incidentally, I think this balance is something that was more commonplace in previous generations, but we've lost to some extent because of very vocal minorities on both fringes. Or, if it's still as common, it's not the approach the more vocal people are taking.)

The lectures are mainly focused around Knuth's 3:16 project, and if you can make it through this book without ever getting the urge to run out and pick up that book as well, then you probably weren't paying attention. The 3:16 project is essentially Knuth's attempt to apply random sampling in order to deepen his understanding of the Bible. Knuth studied, read commentary on, and actually re-translated each "3:16" verse in the Bible.

There's a lot of discussion as to just why he chose this scheme, as well as how he got over some of the hurdles it presented (not every book in the Bible has 3 chapters, for example). Knuth is very clear that there's nothing magic about this approach--he's not looking for some sort of Bible code. Nor does he think this thorough approach over a small sample gives him a thorough knowledge of the Bible. Rather, by thoroughly studying a small portion of the Bible, he believes it's possible to actually gain a deeper understanding of the whole work. You can't read every commentary and every translation for every single verse, but you can be pretty thorough if you limit your study to 59 of them.

The lecture on how he re-translated each of the 59 verses from the original Hebrew and Greek is the most fascinating, and well worth reading even if you're not interested in any of the other lectures. What makes it so interesting is he doesn't know Greek or Hebrew, and yet he was able to pull this off by using a lot of study aids, such as a complete concordance and an in-line English/Hebrew translation. The point, of course, is not to create yet another translation of the Bible--the point is getting past the familiar readings of the text and actually wrestling with what it's really trying to say.

On the flip side, Knuth takes the next lecture to cover aesthetics, in which he goes through the different renderings of the verses he used to illustrate the book that came out of his 3:16 project. There's a lot of discussion on how the calligraphy evokes an image of the verse's content, and interesting stories from the editing process.

The final lecture in the series is more in line with the traditional understanding of how computing relates to religion. (Or rather, how math and logic relate to religion.) It's one thing to argue about things like free will or the omnipotence of God or infinity in generalized terms. It's a completely different thing to frame those distinctions in mathematical terms. Knuth argues that distinctions between the infinite and the finite are minor (in terms of human understanding, at least) when considering huge finite numbers or decision trees. Again, this is where the math just goes over my head, but from what I can understand, it's a very different perspective on these problems, and you'll come away feeling just how simplistic the stereotypical approaches to debating these questions can be.

Anyway, it's a good read if you've got some interest in religion and computers. It's not a particularly life-changing read, nor is it going to give you a completely new perspective. But it is a fascinating approach to Christianity, and it will make you think.

Comments