5 ms·

Learning Algorithms

60 points by paedubucher 3y ago

rg111 3y ago
The author would do himself a great service if he chooses some other resource for learning algorithms. CLRS isn't bad. It is meant for a reference, and it also assumes that you are taking the course in college with an instructor teaching you. There are other great resources out there. Two I recommend are: - Algorithms, Coursera (Sedgewick, Wayne) - The Algorithm Design Manual (Skiena) I have recommended self-taught programmers in the past, which all of them liked: - A Common Sense Guide to Data Structures and Algorithms, PragProg (Jay Wengrow) - Algorithms Illumited, book or Coursera MOOC (Tim Rougharden)
- sdenton4 3y ago
  Skiena is great. My usual learn-a-new-language routine is speed-running the first 25ish Project Euler problems. They cover a lot of three usual language basics. https://projecteuler.net/ https://projecteuler.net/
  - sn9 3y ago
    Skiena posts recent semesters of lectures on Youtube as playlists, as well. Going through them at 1.25-1.5x speed, a lecture per night or every few days, is extremely helpful.
- speakspokespok 3y ago
  I always recommend 'A Common Sense Guide to Data Structures and Algorithms' as well but I'd like to call out the 2nd edition has expanded the recursion chapters and in general the end of chapter assignments are better. The 2nd edition was born from several years of teaching material from the first edition.
- paedubucher 3y ago
  One issue I have with CLRS is that array indices usually start at 1, but sometimes they use 0, because it works better in that particular case. Some functions I used in Erlang (lists:sublist/3, lists:nth/2) use 1-based indices, too. I thought that 0-based array indices (and exclusive upper bounds) were quite a done deal in computer science (https://www.cs.utexas.edu/users/EWD/transcriptions/EWD08xx/EWD831.html https://www.cs.utexas.edu/users/EWD/transcriptions/EWD08xx/E...).
- PartiallyTyped 3y ago
  CLRS and Tardos && Kleinberg are great if you spend a whole year on the book and have mathematical background. The latter is great for algorithms because it groups everything by concepts, ie greedy, DP and then network flows.
BossingAround 3y ago
I've recently been trying to learn algorithms as a self-taught software engineer. I've made it quite far without much deeper knowledge of the area (I do have some basic overview of course), but it has always been on my mind that I should to study up. I've tried CLRS, but oh boy it's a) a slog to go through, b) very sparse with explanations. You might say that the whole book is one explanation after another, and you'd be technically right, but there's nothing more frustrating than an overly complex explanation that might be absolutely precise yet complex to understand. I also often feel like the language is needlessly complex. On top of that, for example, the mathematical explanations often skip steps that are obvious to authors. You might think it's reasonable, but to a struggling newcomer, this is a huge drag. Unrelated to algorithms (but not totally), I am now studying Discrete Mathematics for Comp Sci [0] since I felt like my math knowledge (or lack thereof) is holding me back in reasoning about algorithms. Even those courses have a fair share of skipping obvious steps in mathematic explanations, but it's doable so far (especially with tools like ChatGPT and Wolfram Alpha to check my understanding). I'm just trying to be a better engineer, and I'm having a really hard time trying to find a suitable resource to learn from. [0] https://www.coursera.org/specializations/discrete-mathematics https://www.coursera.org/specializations/discrete-mathematic...
- trealira 3y ago
  I'm still in university, but I took discrete math at my university last semester and took data structures over the summer between semesters. My data structures course mainly was graded from Zybooks assignments. I think it's actually a pretty good and understandable resource. It was this: https://www.zybooks.com/catalog/data-structures-essentials/ https://www.zybooks.com/catalog/data-structures-essentials/ Anyway, maybe after you study discrete math from Coursera, that Zybook could help. It has some basics about big-O notation and determining the big-O of an algorithm (and calculating the number of operations it requires), but after that, it's all explanations of data structures accompanied by many animations and pictures. It may be tedious for someone who already knows how e.g. red black trees works, but it helped me understand them, as well as unbalanced BSTs, AVL trees, B-trees, heaps, and graphs. The stacks and queues were tedious for me, because I already know them. I do think having taken discrete math helped me; I still flipped through CRLS during the course on the side, and their explanations were less impenetrable because I know the basics of mathematical proofs now. There's also a book on data structures that went over hash functions (it's out of print, but I found a PDF online). It was also helpful for me. I'll reply to this comment with it when I get home.
colesantiago 3y ago
Are there any startups doing or requiring complex or exotic algorithms these days to solve unique engineering problems? Would really love to know to see different and various use cases.
zelphirkalt 3y ago
Can anyone recommend a book, that deals with purely functional algorithms? I already have "Purely Functional Data Structures" by Okasaki, which I began working through, but still have a long way to go. However, it is about data structures, not about algorithms, strictly speaking. Perhaps there is a book for algorithms that does it the functional way? Probably a lot of it comes down to naive versions of algorithms using functional data structures and then some changes in the algorithms due to optimizing for functional data structures?
- stathibus 3y ago
  There is a good reason this book doesn't exist.
  - the_sleaze9 3y ago
    Do tell
  - mochaki 3y ago
    and what might those reasons be?
- paedubucher 3y ago
  I haven't read it, nor do I own it, but "Algorithm Design with Haskell" might fit: https://www.cambridge.org/core/books/algorithm-design-with-haskell/824BE0319E3762CE8BA5B1D91EEA3F52 https://www.cambridge.org/core/books/algorithm-design-with-h...
anoy8888 3y ago
Don’t forget to try out Pharo which is modern small talk
- paedubucher 3y ago
  Maybe for the next project; together with Haskell.

60 points by paedubucher 3y ago

The author would do himself a great service if he chooses some other resource for learning algorithms. CLRS isn't bad. It is meant for a reference, and it also assumes that you are taking the course in college with an instructor teaching you. There are other great resources out there. Two I recommend are: - Algorithms, Coursera (Sedgewick, Wayne) - The Algorithm Design Manual (Skiena) I have recommended self-taught programmers in the past, which all of them liked: - A Common Sense Guide to Data Structures and Algorithms, PragProg (Jay Wengrow) - Algorithms Illumited, book or Coursera MOOC (Tim Rougharden)

sdenton4 3y ago
Skiena is great. My usual learn-a-new-language routine is speed-running the first 25ish Project Euler problems. They cover a lot of three usual language basics. https://projecteuler.net/ https://projecteuler.net/
- sn9 3y ago
  Skiena posts recent semesters of lectures on Youtube as playlists, as well. Going through them at 1.25-1.5x speed, a lecture per night or every few days, is extremely helpful.
speakspokespok 3y ago
I always recommend 'A Common Sense Guide to Data Structures and Algorithms' as well but I'd like to call out the 2nd edition has expanded the recursion chapters and in general the end of chapter assignments are better. The 2nd edition was born from several years of teaching material from the first edition.
paedubucher 3y ago
One issue I have with CLRS is that array indices usually start at 1, but sometimes they use 0, because it works better in that particular case. Some functions I used in Erlang (lists:sublist/3, lists:nth/2) use 1-based indices, too. I thought that 0-based array indices (and exclusive upper bounds) were quite a done deal in computer science (https://www.cs.utexas.edu/users/EWD/transcriptions/EWD08xx/EWD831.html https://www.cs.utexas.edu/users/EWD/transcriptions/EWD08xx/E...).
PartiallyTyped 3y ago
CLRS and Tardos && Kleinberg are great if you spend a whole year on the book and have mathematical background. The latter is great for algorithms because it groups everything by concepts, ie greedy, DP and then network flows.

BossingAround 3y ago

I've recently been trying to learn algorithms as a self-taught software engineer. I've made it quite far without much deeper knowledge of the area (I do have some basic overview of course), but it has always been on my mind that I should to study up. I've tried CLRS, but oh boy it's a) a slog to go through, b) very sparse with explanations. You might say that the whole book is one explanation after another, and you'd be technically right, but there's nothing more frustrating than an overly complex explanation that might be absolutely precise yet complex to understand. I also often feel like the language is needlessly complex. On top of that, for example, the mathematical explanations often skip steps that are obvious to authors. You might think it's reasonable, but to a struggling newcomer, this is a huge drag. Unrelated to algorithms (but not totally), I am now studying Discrete Mathematics for Comp Sci [0] since I felt like my math knowledge (or lack thereof) is holding me back in reasoning about algorithms. Even those courses have a fair share of skipping obvious steps in mathematic explanations, but it's doable so far (especially with tools like ChatGPT and Wolfram Alpha to check my understanding). I'm just trying to be a better engineer, and I'm having a really hard time trying to find a suitable resource to learn from. [0] https://www.coursera.org/specializations/discrete-mathematics https://www.coursera.org/specializations/discrete-mathematic...

trealira 3y ago
I'm still in university, but I took discrete math at my university last semester and took data structures over the summer between semesters. My data structures course mainly was graded from Zybooks assignments. I think it's actually a pretty good and understandable resource. It was this: https://www.zybooks.com/catalog/data-structures-essentials/ https://www.zybooks.com/catalog/data-structures-essentials/ Anyway, maybe after you study discrete math from Coursera, that Zybook could help. It has some basics about big-O notation and determining the big-O of an algorithm (and calculating the number of operations it requires), but after that, it's all explanations of data structures accompanied by many animations and pictures. It may be tedious for someone who already knows how e.g. red black trees works, but it helped me understand them, as well as unbalanced BSTs, AVL trees, B-trees, heaps, and graphs. The stacks and queues were tedious for me, because I already know them. I do think having taken discrete math helped me; I still flipped through CRLS during the course on the side, and their explanations were less impenetrable because I know the basics of mathematical proofs now. There's also a book on data structures that went over hash functions (it's out of print, but I found a PDF online). It was also helpful for me. I'll reply to this comment with it when I get home.

colesantiago 3y ago

Are there any startups doing or requiring complex or exotic algorithms these days to solve unique engineering problems? Would really love to know to see different and various use cases.

zelphirkalt 3y ago

Can anyone recommend a book, that deals with purely functional algorithms? I already have "Purely Functional Data Structures" by Okasaki, which I began working through, but still have a long way to go. However, it is about data structures, not about algorithms, strictly speaking. Perhaps there is a book for algorithms that does it the functional way? Probably a lot of it comes down to naive versions of algorithms using functional data structures and then some changes in the algorithms due to optimizing for functional data structures?

stathibus 3y ago
There is a good reason this book doesn't exist.
- the_sleaze9 3y ago
  Do tell
- mochaki 3y ago
  and what might those reasons be?
paedubucher 3y ago
I haven't read it, nor do I own it, but "Algorithm Design with Haskell" might fit: https://www.cambridge.org/core/books/algorithm-design-with-haskell/824BE0319E3762CE8BA5B1D91EEA3F52 https://www.cambridge.org/core/books/algorithm-design-with-h...

anoy8888 3y ago

Don’t forget to try out Pharo which is modern small talk

paedubucher 3y ago
Maybe for the next project; together with Haskell.