Ask HN: Is Knuth's TAOCP worth the time and effort? • Hacker News Board

Re: Ask HN: Is Knuth's TAOCP worth the time and effort?

#101

Post by microcolonel » Thu, Jan 14, 2016, 7:07 AM UTC

apetresc wrote: post #25

It's not useful as a reference, but it is full of delight – though only if you actually do the problems. Many of them are chains of inquiry that build on each other and culminate in some result that gives you so much intuition about whatever it was they were dealing with. If you're in the Toronto region, I actually run a reading group for these books: http://www.meetup.com/Knuth-Reading-Group-Art-of-Computer-Pr... .…

I keep neglecting to come, mostly out of shame; frankly I find the texts troublingly dense and don't want to feel weak in public.

Re: Ask HN: Is Knuth's TAOCP worth the time and effort?

#102

Post by zvrba » Thu, Jan 14, 2016, 7:15 AM UTC

JackFr wrote: post #46

Retric wrote: post #11

pipeline stalls, designing for cache performance, branch prediction, multithreading, etc. Those are issues for a subset of computer hardware. Embedded computing is still huge and has drastically different performance requirements. TAOCP is designed to stand the test of time, not focus on micro-optimizations that generally have limited performance impact and are a waste of time 99% of the time.

> TAOCP is designed to stand the test of time, not focus on micro-optimizations . . . Hmmm. My edition of "Sorting and Searching" has a section devoted to tape-sorting.

He plans to remove that section in the next edition of the book :( I'd argue the section is very relevant since today's RAM works fastest when accessed sequentially.

Re: Ask HN: Is Knuth's TAOCP worth the time and effort?

#103

Post by zvrba » Thu, Jan 14, 2016, 7:24 AM UTC

When I'm bored I pick a random section and read it. I was really fascinated by the section on sorting networks for example.

The real gems are the exercises and answers to them; there's at least as much useful/applicable stuff as in the main text. I tried to solve them, but found myself unable to proceed with them if difficulty was > 20, even if I understood the concepts. I have no idea on how to attack them. (Even non-mathematical ones.)

IMHO, TAOCP needs "Volume 0" which would teach you problem solving in this particular domain. (No, "Concrete mathematics" is not it. That book is, IMHO, awful: many problems depend on having tricks up your sleeve.)

Any tips on how to approach exercises which seemingly don't have a lot to do with the preceding text?

Re: Ask HN: Is Knuth's TAOCP worth the time and effort?

#104

Post by lokedhs » Thu, Jan 14, 2016, 7:29 AM UTC

pajop wrote: post #89

it's worth reading just for one of the best book in-jokes ever - see the introductory part "Notes on the Exercises" of Volume 2 "Seminumerical Algorithms" 3. [M50] Prove that when n is an integer, n > 2, the equation xn + yn = zn has no solution in positive integers x, y, z. Knuth's book was published in 1973. The initial solution to this exercise was solved by a then Princeton professor in 1993 and was finalized by…

If I recall correctly, newer editions of the book has lowered the difficulty of this (M49?) as it has been solved.

Re: Ask HN: Is Knuth's TAOCP worth the time and effort?

#105

Post by laichzeit0 » Thu, Jan 14, 2016, 7:41 AM UTC

marvy wrote: post #73

davidw wrote: post #56

Earlier quoted context omitted.

TAOCP has not been written, several volumes of it have. It remains to be seen whether Knuth will even be able to finish what he set out to do, let alone include new things. > faster than can be documented. By one person , I wrote. Here's one that I think is interesting, difficult and evolving: distributed systems. This shows interesting work happening with version vectors, for instance, over the past 15 years: https:…

No need to worry about that one. Looking at Wikipedia's list of volume's, looks like distributed systems are not being covered. But I agree with you anyway. Let's quote the man himself (taken from his website): "As I continue to write Volumes 4 and 5, I'll need to refer to topics that belong logically in Volumes 1--3 but weren't invented yet when I wrote those books". He plans to complete Volume 5 by 2025. Tune in te…

Knuth is currently 78 years old. It would be interesting to see if he lives to 88, never mind still have the intellectual and or physical capacity to complete such a work.

I wish him the best of lucky and sincerely hope he pulls it off.

Re: Ask HN: Is Knuth's TAOCP worth the time and effort?

#106

Post by throwaway2048 » Thu, Jan 14, 2016, 7:43 AM UTC

kamaal wrote: post #62

>>was it worth it? You are not supposed to read about any algorithm. There is nothing special about knowing how some sorting algorithm works, or any algorithm for that matter. That knowledge is no better than knowing how some war in history happened. Or remembering arcane trivial about things. Knowing how to sort a list a dozen ways, no big deal. Unless you discovered the algorithms yourself, there is nothing special…

It virtually all programming, even in stuff with tight performance constraints and high performance demands you are utilizing known and studied algorithms.

Re: Ask HN: Is Knuth's TAOCP worth the time and effort?

#107

Post by cgag » Thu, Jan 14, 2016, 7:51 AM UTC

I haven't read TAOCP, but for those things you mentioned, Computer Systems a Programmer's Perspective is great, especially if you're thinking the way you seem to be about how you can exploit these things rather than how they work in depth.

Re: Ask HN: Is Knuth's TAOCP worth the time and effort?

#108

Post by eyan » Thu, Jan 14, 2016, 8:20 AM UTC

> So for those who are practical programmers and have gone down the Knuth TAOCP rabbit hole

There's the rub. TAOCP is Computer Science. "Practical programmers" aren't even engineers. So yes. The books for your intended purpose aren't worth anything.

It will have worth when you are writing something 'serious' like compilers, kernels, or complicated games. CRUD apps, nah.

Re: Ask HN: Is Knuth's TAOCP worth the time and effort?

#109

Post by ajarmst » Thu, Jan 14, 2016, 8:28 AM UTC

I sometimes think of Computer Science (in the classical theoretical sense---and Knuth is a towering figure in that domain) as having a similar relationship with programming as that of Mathematics to Engineering. A basic foundation in the theoretical background is very helpful, possibly essential, to professional work, but a deep understanding of theory is unnecessary. That said, those engineers that have a deeper mathematical foundation can realize important benefits to the quality of their work. Similarly, a rich immersion in abstract computer science is not necessary to programming, but its presence can be of great value. Being aware that a given programming problem maps to a well-understood class of algorithms with known performance and correctness enables elegant solutions.

Re: Ask HN: Is Knuth's TAOCP worth the time and effort?

#110

Post by ajarmst » Thu, Jan 14, 2016, 8:35 AM UTC

ajarmst wrote: post #109

I sometimes think of Computer Science (in the classical theoretical sense---and Knuth is a towering figure in that domain) as having a similar relationship with programming as that of Mathematics to Engineering. A basic foundation in the theoretical background is very helpful, possibly essential, to professional work, but a deep understanding of theory is unnecessary. That said, those engineers that have a deeper mat…

I should note that, while TAOCP is absurdly important in CompSci, it's probably not a good starting point. I'd recommend Corman's book on algorithms first, and maybe getting your feet wet with Knuth's Concrete Mathematics ( if only as a more gentle introduction to his style).