Classic Algorithms Implemented in JavaScript | Hacker News Reader

Classic Algorithms Implemented in JavaScript

github.com

249 pointstrekhleb34 comments

The implementation of a hash table uses a hash table, which feels like cheating to me.

Also I suspect that adding together the ascii values of the characters in the key is not necessarily the best hash function.

the keywords "ill" and "duh" will cause collision. I know this because I ran into this issue trying to be clever on leetcode's anagram grouping problem.

Using what hash mechanism? I tested with SHA512 and got different results:

* ill - 2ae60a99f87408af77066079879865d907201ff44e7b1f26d911afcd1097c4c32162b97546c404d3a305396698ea530d3c3690adb7dcbc91ec685837f895a37b

* duh - e44e6d93cd165ec8cdeaa632abe1c35d9a71533cc68072e100c5307ea3be17bc4f10e258785bef77d31664be0f1c3f427528c7450d52082a90c4ee86c7012184

Adding their ASCII together, they both yield 321.

What jovial said. Adding ascii yields the same integer for ill and duh.

Yeah, there are 26^n length n strings of lowercase alphabetical characters, but only ~26*n different hashes with that function.

> Also I suspect that adding together the ascii values of the characters in the key is not necessarily the best hash function.

Still better than using the length of the string as your hash value. Not my proudest moment but sometimes, you want to see the world burn. That and I forgot to do my hw and class was in a few minutes.

What I would wish for is a list of common problems and how certain algorithms and data structures solve them.

I recommend Skiena's "The Algorithm Design Manual", it also includes some fun war stories of real world problems he was called in to solve, and the algorithms he used to solve them.

Well its not a list per se but many of the interview prep sites tag their problems with the appropriate algorithm/data structure for the job. As someone who has his first SV phone technical interview on Monday and has done ~30 problems in the last two weeks, I can say after solving about 2-3 of them over a number of hours or days you start to recognize which algorithm will get the job done from the problem descriptions.

Interesting :D

Would be nice if you could post a list of your resources :)

Not OP, but I used leetcode extensively while prepping for interviews and it worked out well for me. Like she/he said, just spend a few hours a day doing these problems; I even went back to problems I'd previously solved to see if I could write out the correct code in one shot.

Not OP, but a resource commonly suggested is "Cracking the Coding interview"

Would recommend checking out the algorithm design manual if you're looking for a book with that sort of approach.

Edit: just noticed this was already recommended below, take this as another strong endorsement!

This is amazing! I’ve forked this with plans to translate all of it to TypeScript, which really just means adding <T> about 10 times per file, erasing the java doc comments and replacing them with actual parameter types. Still worth it!

As poor as I did in my Algorithms class last semester, this would have been very helpful to reference. There's definitely a gap between the concept and implementation of CS concepts, being able to see an efficient implementation of something like Heap Sort really helps teach how it works.

Super job! As someone currently studying this stuff, this is a gold mine. It's hard to interview in this field if you didn't get a CS degree, and this sort of gesture is forever appreciated.

It’s hard to interview in this field even with a CS degree, unfortunately. The problem is there are too many people interviewing, with and without CS degrees.

I haven't run into the "too many people interviewing" problem, though I am not the interviewer.

My main issue is definitely the whiteboard portion of the competition. Complaining about those whiteboard skills being of negligible daily use on the job is getting me nowhere: the only solution is to study up. My strongest language is JS, so this repo is really, really nice.

I'm impressed by the commenting. This is much higher quality stuff than I expected!

Great resource, but I couldn't help laughing at the plot with asymptotic complexity classes categorized as "horrible, bad, fair, good, excellent", as if it's independent of the problem.

It would be cool if these had tests around them where it made sense.

Already there:

https://github.com/trekhleb/javascript-algorithms/blob/maste...

Beautiful. Apparently, I was not paying attention.

its pretty interesting to see the steep curve for how many stars this repo has collected in such a short amount of time

http://www.timqian.com/star-history/#trekhleb/javascript-alg...

It'd be nice if they made the distinction between functional/immutable data structures and mutable ones.

Lets pretend JS arrays had no queue api, why whouldn't I use native arrays to implement a queue?

Doesn't look like a utility belt but a reference. If I want to know how queue is implemented by example, using the native queue methods wouldn't help whatsoever.

Cool list but I would like to see them being implemented using FP approach when possible.

You can always fork it and make a PR

Love the beginner/advanced tags. Also, just generally well-organized. This is great.

Nicely done git web. Love these things done "right"

No FizzBuzz?

Not really the kind of thing people usually mean when they talk about "algorithms."

Reply on news.ycombinator.com