8 ms·

This book is often recommended, but I would dearly love to know how many of its recommenders have actually read it.

by shanusmagnus 10y ago

hossbeast 10y ago
There's an algorithm for that ...
- nekopa 10y ago
  It ends up in an infinite loop for me :(
Hernanpm 10y ago
I rather recommend this book end to end by is a good to own it, when I use to compete in programming competitions from time to time there were interesting problems that have an algorithm explained in this book but not in others it is a good reference overall.
Koshkin 10y ago
Well, it is a bible, after all.
nekopa 10y ago
I have read the first 50 pages about a 100 times. Does that count? P.S. I have never recommended it to anyone though, I don't recommend books I haven't read all the way through. I will get through TAOCP one day, (it's now on my retirement bucket list, along with learning Sanskrit) but I first need to read through a bunch of Math books forst, so I can get past the 50 page hump.
- kyberias 10y ago
  I wonder, what might be a suitable set of books to read before TAOCP? Might this one suffice: Concrete Mathematics: A Foundation for Computer Science (Knuth et.al.)
  - gajjanag 10y ago
    I think yes, but then I have not read TAOCP beyond the first chapter. What I can say is that "Concrete Mathematics" is a terrific book. It is the best general "math for algorithms" book I have seen. Its exercises are wonderfully picked, with problems at all levels of difficulty and classified neatly into categories. It also discusses topics that for strange reasons are not as widely known as one would expect, even though they are extremely elegant and quite useful. A concrete example of this is the Euler-Maclaurin formula relating sums and integrals.
- hvmonk 10y ago
  off the topic, but what made you curious about Sanskrit?
  - vram22 10y ago
    I'm not the OP, but can comment about Sanskrit a bit. (I had studied Sanskrit in school.) [Edit:] As a third language, apart from English and Hindi. This is mostly all IMO, except where specified otherwise: - it's an interesting language, because it is the ancestor of many Indian languages (except the languages of the southern Indian states - Tamil, Telugu, Malayalam and Kannada - and their dialects - and except the languages of the North-Eastern states beyond Assam, maybe - not sure on this one). In fact, even those languages (southern Indian at least) have some Sanskrit-derived words; I think the term is loanwords. - it's grammar rules are very regular (compared to many other languages, I've heard), which makes it less of a chore and easier to remember and learn; - (probably many) parts of the language are sort of mathematical / logical in the sense that some of the rules build upon other rules, so you can use your knowledge of the latter to understand the former; e.g. a part of the language called sandhi, which gives the rules for how to combine smaller words/sounds into larger ones. Once you learn those rules (and there are only a handful), a) you can use them both to create new words according to those sandhi rules, by combining existing words (and those new words would be legal Sanskrit, BTW - poets and prose writers did it all the time), and b) can use them to figure out what compound words mean (that you have newly come across), i.e. the reverse process of a). So in that and maybe some other aspects its a bit like a programming language [1], where you build up higher abstractions from lower-level ones. [Edit:] It's my guess that people who like languages like Lisp, Haskell, Clojure, etc. - that are built upon mathematical principles and so on, might find Sanskrit interesting to learn. - [1] I've also read a few times that due to the mentioned regular qualities of the grammar (and also, I think, due to the unambiguous nature of the sounds in it - each letter or small letter grouping can be (legally) pronounced in only one way), CS and Sanskrit-knowing people have done research on using Sanskrit in computer research in various ways (programming languages (research), AI? No idea about the success or not of those efforts, though.
    - Koshkin 10y ago
      Indeed, these are all good qualities of a fine language. It's just that I am not sure if it is those qualities what attracts people and makes them want to learn it. There also must be something else.
marrs 10y ago
I'm in the process of reading Vol 1 atm. It's densely packed but really well presented so far. I should caveat that I'm still near the beginning (learning MIX) so I can't claim that it will remain that way, but so far I'm really enjoying it
ska 10y ago
It's a wonderful read, and a good education. On the other hand, it isn't a very practical solution to most people who are thinking "I'm not sure what algorithm to use here".
importantbrian 10y ago
I don't think the point of this book is to read it cover to cover, or in this case cover to cover to cover to cover...since it's going on 4 volumes now. I think it's meant more as a reference. Pull it off the shelf and read the relevant section when you need it.

This book is often recommended, but I would dearly love to know how many of its recommenders have actually read it.

hossbeast 10y ago

There's an algorithm for that ...

nekopa 10y ago
It ends up in an infinite loop for me :(

Hernanpm 10y ago

I rather recommend this book end to end by is a good to own it, when I use to compete in programming competitions from time to time there were interesting problems that have an algorithm explained in this book but not in others it is a good reference overall.

Koshkin 10y ago

Well, it is a bible, after all.

nekopa 10y ago

I have read the first 50 pages about a 100 times. Does that count? P.S. I have never recommended it to anyone though, I don't recommend books I haven't read all the way through. I will get through TAOCP one day, (it's now on my retirement bucket list, along with learning Sanskrit) but I first need to read through a bunch of Math books forst, so I can get past the 50 page hump.

kyberias 10y ago
I wonder, what might be a suitable set of books to read before TAOCP? Might this one suffice: Concrete Mathematics: A Foundation for Computer Science (Knuth et.al.)
- gajjanag 10y ago
  I think yes, but then I have not read TAOCP beyond the first chapter. What I can say is that "Concrete Mathematics" is a terrific book. It is the best general "math for algorithms" book I have seen. Its exercises are wonderfully picked, with problems at all levels of difficulty and classified neatly into categories. It also discusses topics that for strange reasons are not as widely known as one would expect, even though they are extremely elegant and quite useful. A concrete example of this is the Euler-Maclaurin formula relating sums and integrals.
hvmonk 10y ago
off the topic, but what made you curious about Sanskrit?
- vram22 10y ago
  I'm not the OP, but can comment about Sanskrit a bit. (I had studied Sanskrit in school.) [Edit:] As a third language, apart from English and Hindi. This is mostly all IMO, except where specified otherwise: - it's an interesting language, because it is the ancestor of many Indian languages (except the languages of the southern Indian states - Tamil, Telugu, Malayalam and Kannada - and their dialects - and except the languages of the North-Eastern states beyond Assam, maybe - not sure on this one). In fact, even those languages (southern Indian at least) have some Sanskrit-derived words; I think the term is loanwords. - it's grammar rules are very regular (compared to many other languages, I've heard), which makes it less of a chore and easier to remember and learn; - (probably many) parts of the language are sort of mathematical / logical in the sense that some of the rules build upon other rules, so you can use your knowledge of the latter to understand the former; e.g. a part of the language called sandhi, which gives the rules for how to combine smaller words/sounds into larger ones. Once you learn those rules (and there are only a handful), a) you can use them both to create new words according to those sandhi rules, by combining existing words (and those new words would be legal Sanskrit, BTW - poets and prose writers did it all the time), and b) can use them to figure out what compound words mean (that you have newly come across), i.e. the reverse process of a). So in that and maybe some other aspects its a bit like a programming language [1], where you build up higher abstractions from lower-level ones. [Edit:] It's my guess that people who like languages like Lisp, Haskell, Clojure, etc. - that are built upon mathematical principles and so on, might find Sanskrit interesting to learn. - [1] I've also read a few times that due to the mentioned regular qualities of the grammar (and also, I think, due to the unambiguous nature of the sounds in it - each letter or small letter grouping can be (legally) pronounced in only one way), CS and Sanskrit-knowing people have done research on using Sanskrit in computer research in various ways (programming languages (research), AI? No idea about the success or not of those efforts, though.
  - Koshkin 10y ago
    Indeed, these are all good qualities of a fine language. It's just that I am not sure if it is those qualities what attracts people and makes them want to learn it. There also must be something else.

marrs 10y ago

I'm in the process of reading Vol 1 atm. It's densely packed but really well presented so far. I should caveat that I'm still near the beginning (learning MIX) so I can't claim that it will remain that way, but so far I'm really enjoying it

ska 10y ago

It's a wonderful read, and a good education. On the other hand, it isn't a very practical solution to most people who are thinking "I'm not sure what algorithm to use here".

importantbrian 10y ago

I don't think the point of this book is to read it cover to cover, or in this case cover to cover to cover to cover...since it's going on 4 volumes now. I think it's meant more as a reference. Pull it off the shelf and read the relevant section when you need it.