Hacker News Reader

Comment by h2odragon | Hacker News Reader

Knuth's "art of Computer Programming", especially and specifically for the way it communicates things. I'm quite weak when it comes to reading advanced maths, but the code communicated to me, and taught me most of what i ever learned about how to read the weird squiggles.

View on HN

This book is recommended (and heavily upvoted) on every such thread about good software books but I would very much doubt more than 5% have read more than a few chapters, probably less than 1% have read one book in the series and probably less than 0.1% have read the entire series.

This is one of those series that everyone thinks they should read but few actually do.

I read all 3 volumes obsessively in high school. I don't claim to have understood much of them, but every page passed before my eyes, and I managed to make sense of the easier parts, I think. One thing that stuck with me forever was the exercises near the beginning of volume 3, where you are given various "data processing"-ish questions that you have to solve basically with sorting and merging operations on a computer without much memory. I used those methods over the years to do stuff with the GNU sort utility and simple shell scripts that would have taken much longer to do with e.g. SQL queries. Being fairly SQL-clueless finally caught up with me a few years ago, but it was a good run.

they make great bathroom books; pick 'em up and open a random page and there's something cool, more often than not.

You can make the same arguments about "do people really read their Bibles?" and the answer is it don't matter: if there's any benefit to be had from the book at all, it's much more likely to accrue if you have a copy.

The book is certainly approachable to read and understand. Especially with some amount of background math. I think the book is hyped so much because people don't have the right background and/or find it difficult at the beginning and give up. Same goes for SICP. And people think the books are really good just because it's hard. It's not, it's approachable given the right effort.

I think it's more a question of how many people fully understand it once having read it.

I recently started trying to understand non-cryptographic hash functions and pretty much every text out there ends up referencing Knuth without really explaining it deeply. Hashing (not hash tables, just hashing) is a subject that pretty much every book seems to ignore and just go with the "here, use this hashing thing because Knuth said so [TAOCP volume 3]". I had to buy it.

That said, I also read other chapters when I had the need to understand specific things.

Reply on news.ycombinator.com