Re:Red-Black Trees in a Functional Setting (Okasaki, 1993)
by Xichekolas on Tuesday March 16, 2010 @02:14PM
(#1196396)
Good point. Judging from the table of contents, he added a chapter up front on various types of trees, tacked on Haskell source in the appendix, and mushed all the conclusions together.
But I'm a poor grad student, so I'll read the free dissertation over the $75 book. ;)
In the same directory, Kahrs develops other versions with stronger constraints coming from the type system, but I never got far enough in to Haskell back then to understand them.
Re:Red-Black Trees in a Functional Setting (Okasaki, 1993)
by kapuzineralex on Tuesday March 16, 2010 @03:35AM
(#1195239)
I find it particularly interesting that he breaks down the number of neccessary cases to only 4 compared to the standard explanation found everywhere else.
Re:Red-Black Trees in a Functional Setting (Okasaki, 1993)
by thirdstation on Tuesday March 16, 2010 @10:58AM
(#1195808)
I took his Programming Languages and Translators class when he was at Columbia. It was, hands-down, the best class I've ever taken. His enthusiasm for the subject was evident. The aspect about the class I liked best, and which I try to emulate, it that he rarely ever directly answered a question. He would, with incredible patience, ask you questions in-turn to help lead you to the solution or definition. I always found that impressive from a graduate-level professor.
I think I still have his ML implementation of Red-Black trees he gave the class.
http://www.cs.cmu.edu/~rwh/theses/okasaki.pdf
I'd highly recommend it. It's hard to find a dissertation that reads like a novel, but he somehow managed to accomplish such a feat.
But I'm a poor grad student, so I'll read the free dissertation over the $75 book. ;)
There is an implementation that is pretty faithful to Okasaki's paper here: http://www.cs.kent.ac.uk/people/staff/smk/redblack/Untyped.h... . I used as a reference once because it also implements delete.
In the same directory, Kahrs develops other versions with stronger constraints coming from the type system, but I never got far enough in to Haskell back then to understand them.
http://matt.might.net/articles/implementation-of-immutable-p...
I think I still have his ML implementation of Red-Black trees he gave the class.