5 ms·
Ten Years of Purely Functional Data Structures (2008)
- oinksoft 14y ago(2008)
- fyi80 14y ago"Five Years of '10 Years of Purely Functional Data Structures'"
- andrewcooke 14y agoit mentions me! my one claim to fame... here's the review http://developers.slashdot.org/story/04/02/19/2257203/purely-functional-data-structures http://developers.slashdot.org/story/04/02/19/2257203/purely...
- DanWaterworth 14y agoI wonder what other interesting books are out there that just need someone to review them on slashdot.
- juliangamble 14y agoThese ideas are one of the foundations of the Clojure language. Rich Hickey mentions in one of his talks that this book got him quite excited. You can see reference to it in Rich Hickey's Clojure bookshelf on Amazon http://www.amazon.com/Clojure-Bookshelf/lm/R3LG3ZBZS4GCTH http://www.amazon.com/Clojure-Bookshelf/lm/R3LG3ZBZS4GCTH
- kwindla 14y agoThat's a great bookshelf list; lots of good stuff that isn't widely referenced. "The Art of the Metaobject Protocol" is an under-appreciated classic, for example.
- vowelless 14y agoHis thesis is quite useful: http://www.cs.cmu.edu/~rwh/theses/okasaki.pdf http://www.cs.cmu.edu/~rwh/theses/okasaki.pdf
- jph 14y agoBrilliant. I can donate $100 to get it updated and as a PDF.
- jfb 14y ago$100 here as well.
- binarycrusader 14y agoWhat is this referring to? I see nothing obvious that mentions a donation or PDF.
- jph 14y agoI'm offering to donate $100 to the author if he would like to update the book as he mentions in his story, and I'd like to get the book as a PDF instead of on paper because of the code samples and searching. This is my personal gift to him because I value his ideas and I hope to learn more. This may also inspire other people here to offer to donate, which will be great if it happens.
- kenjackson 14y agoGreat book. Although I'd like to see an "open source" version when you can get all the code snippets expressed in the language of your choice. When I read the book I wanted Haskell, but now I think I'd like to see it in F# or maybe even C#.
- tikhonj 14y agoIt does have the Haskell versions in the appendix. Besides, depending on your perspective, both SML and Haskell are basically functional pseudocode.
- kenjackson 14y agoThere was something missing from the Haskell stuff, but I can't recall the details -- it's been almost 10 years ago since I read it. But when I am reading a data structures/algorithms book, I don't want to have to deal with code (SML in this case) that I'm not super familiar with. Little things get in the way. With modern software, and the ability to dynamically swap out content, it seems a pity that we don't better tailor content to the reader.
- igravious 14y agoSome questions. Background: I'm coming to the whole world of functional programming via Ruby (via C). Doesn't Clojure have purely functional (immutable) data structures? Whenever this area is talked about I always hear Clojure mentioned. Do other functional languages have libraries (standard or otherwise) of these? Are these data structures implemented in Clojure itself? I kind of see Ruby like a hybrid functional language. Could Ruby be used to implement the data structures in this book, or would that be like a square peg in a round hole? Didn't Ruby recently get official lazy evaluation? Hum, so many questions, so little Google :(
- DanWaterworth 14y ago> I kind of see Ruby like a hybrid functional language. Functions aren't even first class in Ruby (though it does have closures). I certainly wouldn't describe it as a hybrid functional language.
- riffraff 14y agoI'd refrain from saying "functions aren't first class in ruby", unless you qualify the assertion. E.g. you could mean "you can't call functions by just adding () at the end", or "the block syntax only allows passing one function", but these are both just syntax differences. On the other hand, you can create standalone functions, pass them around, introspect them, invoke them, call methods on them and compose them, which is more or less what you can do with integers in python (we can agree integers are first class objects in python, I think) So there may be reason for saying they are "not first class", but you should expand on it.
- DanWaterworth 14y agoYou're right, I should have backed up that assertion. The reason I say that functions aren't first class is that if you create a function: def factorial(n) if n > 0 n * factorial(n - 1) else 1 end end You can't pass it into another function without creating an object for it. ie, using Kernel.method(:factorial). It's the same with blocks, they aren't first class, in order to construct values, you have to construct objects that holds them.
- ludicast 14y agoI took his PLT class at Columbia and TA-ed the class later. Best professor I ever had. I believe he is now at West Point where students won't whine when he hurts their brains. Would like to see the book re-done with Haskell as the base (rather than appendix) because that language is more "purely functional" and seems to finally have the legs it needed back in 2008.