Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

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

news.ycombinator.com

Post reply on HN
Print view View original

141–150 of 192 posts

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

#141

Post by retrac »

Have you read Euclid's Elements? Did you enjoy the proofs? And the general process? The conclusions in the Elements are what any 10th grade geometry student should already know. It's not the conclusions though, but the journey. Some people still swear by Euclid, claiming the Elements helps teach a person how to think.

TAOCP is in that vein. As you have identified, it is not a programming book, and it's not a computer engineering book, and it's not really a college textbook either. It is a computer science book in the old tradition, where computer science is treated as a branch of mathematics.

It won't teach you to program. As another user said, you should probably be using the algorithm your language's standard library provides. Now, if you want to know why a library's author chose a particular algorithm, the trade-offs between the options, and an open-ended and in-depth exploration of such topics, then this is the book for you. The author of that optimized tree search function you're using probably read TAOCP at some point, so you don't have to.

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

#142

Post by adastra22 »

xxs wrote: post #89
adastra22 wrote: post #3

Your milage will vary. I got zero value from the books. They do make a good doorstop though.

book 3, especially sorting and hashing is pretty much contemporary. The linear probe, alone, and deletion shift of 6.4 is more than worth it.

By contemporary you mean relevant? That’s not my complaint. TAOCP is pedagogically bad and the content is better learned from other sources.

People just seem to elevate Knuth to divine status and treat his books like the Bible. Why, idk.

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

#143

Post by adastra22 »

Koshkin wrote: post #120
secondcoming wrote: post #29

Earlier quoted context omitted.

Same. I bought physical copies years ago but ended up giving them away. The main problem is that algorithm examples are specified in a fake machine code. It’s not that useful when things can just be Googled now.

Well, the Turing machine is also “fake.” It is used for some analytical research. Similarly, Knuth uses his MIX for the analysis of algorithms.

It is poorly adapted to that use case though. Either use an abstract machine that captures the essence of the algorithm better, or use a more realistic yet still simple machine (e.g. MIPS or RISC-V). MIX is a middle ground that fails on both fronts.

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

#144

Post by adastra22 »

dragonelite wrote: post #25

Does a christian needs to read the bible to a be christian not really. Same with TAOCP and programmers, but the more you read about CS the more complete of a programmer you become. For example i never thought i would use the basic compiler knowledge i gained during my bachelor. But at one of my first consulting gigs it did helped knowing about abstract syntax tree, where you can add and remove nodes. This knowledge h…

Scientists dont need bibles.

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

#145

Post by JohnFen »

I recommend it because it gives an excellent overview of the fundamentals of the field. Well, honestly, because the series of books were formative for me so I'm a bit biased. I can't overstate the value I got from them. I still use them as reference works to this day. However, I can't say whether or not anyone else would get the same value.

I didn't think of studying them as "effort", though, because they were (and are) extremely interesting to me.

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

#146

Post by hintymad »

I'd recommend it, but I wouldn't recommend reading it cover to cover. It really depends on what you need. Here is what I read selected chapters for:

- For expanding my conceptual understanding of algorithms. I still remember the thrill of learning how the Joseph's problem can be modeled as base-n modulo arithmetics. Or how powerful it is to use generating functions to compute algorithm complexities. And I read the vol 4 for a pleasant study of how we can code back-tracking in an almost mechanical way.

- For historical context. Knuth is amazing at tracing the roots of algorithms, and show how the solutions to a problem evolve over time. You'll get high when seeing how a seemingly impossible problem gets solved in such elegance and ingenuity over generations of attempts by great minds. Check Vol 4 for all the wonders of combinatoric algorithms, for example.

- For thorough study and pure awe. For instance, one probably learns a bit about ordered binary decision diagram (OBDD) to reduce state space by thousands of times in a formal verification system. But what if you want to see all the amazing details of OBDD and therefore getting hours of dopamine high? Vol 4 is the go-to place.

- For reference. At least for a student. I got stuck implementing a fibonacci heap for my project on memory management in my OS class. And Knuth's books came to rescue. I guess professionals would still refer to Knuth for their implementation from time to time.

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

#147

Post by llimos »

dctoedt wrote: post #136
dorfsmay wrote: post #95

Earlier quoted context omitted.

Memento Mori. You should imagine that you are forced to sell it every day. Eventually you'll either sell it (or donate it (your local library might appreciate it)) or start reading it.

> You should imagine that you are forced to sell it every day. Semi-serious question: Which of the following do you mean: A. Every day, you should imagine that you are forced to sell it; or B. You should imagine that you will, every day, be forced to sell it. (I teach law students to spot and fix ambiguities in draft contracts; this would be a nice practice example :-) )

Since it's not possible to sell the same physical item more than once without some intervening event, it's not that ambiguous!

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

#148

Post by dctoedt »

llimos wrote: post #147
dctoedt wrote: post #136

Earlier quoted context omitted.

> You should imagine that you are forced to sell it every day. Semi-serious question: Which of the following do you mean: A. Every day, you should imagine that you are forced to sell it; or B. You should imagine that you will, every day, be forced to sell it. (I teach law students to spot and fix ambiguities in draft contracts; this would be a nice practice example :-) )

Since it's not possible to sell the same physical item more than once without some intervening event, it's not that ambiguous!

https://en.wikipedia.org/wiki/Prometheus ?

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

#149

Post by stolenmerch »

dwheeler wrote: post #90
dmvdoug wrote: post #60

Earlier quoted context omitted.

Principia was an attempt to derive math from logical ground up, motivated by ideas about the foundations of mathematics and logic (wrong ones, it turned out). Knuth is more like A Magical Mystery Tour of Algorithms, wherein Knuth is both tour guide and man behind the curtain.

"Wrong" is too strong. The fundamental bases it used are not generally used today, but it was the first of its kind and inspired much. Many of its details are still fine. If you are interested in the underlying goal of Principia Mathematica, I urge you to check out the Metamath Proof Explorer (MPE): https://us.metamath.org/mpeuni/mmset.html By itself, Metamath doesn't have built-in axioms. MPE uses Metamath to first…

Also check out the Principia Rewrite project, which aims to use the interactive theorem prover Coq to ensure each proof step is a valid step according to Principia’s axioms and that no steps are skipped, even by accident.

https://www.principiarewrite.com/

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

#150

Post by stonogo »

adastra22 wrote: post #139
txs wrote: post #18

Earlier quoted context omitted.

How do you measure the value? I treasure these books, they've brought me so much joy over the years. They've likely not affected my career trajectory much, but they've absolutely enriched my personal life.

There wasn’t anything useful in these books I didn’t already know from an extensive binge into CS theory and practice. There was no new insights either. There was a lot of useless knowledge that wasted time to cover, like how to optimize for his totally made up machine language.

All machine languages are totally made up.
Post reply on HN