Thanks, Kenneth. Actually, I've already forked and used harthur's code several times in the past -- she's written some really cool machine learning stuff for JavaScript that is easy to read. I've used it as the basis for playing around with machine learning in Python and JavaScript, and last year I tried (unsuccessfully, I think) to use d3 to create a nice visualization of her implementation of the k-means algorithm[0].
My visualization turned out poorly (I've had a few suggestions on how to make it better that I need to integrate, and am still open to additional suggestions). However, that obviously had nothing to do with harthur's excellent clusterfck library that I used[1].
Really, when you look over harthur's github contributions, she seems like the ideal open source contributor. She uses permissive licensing, writes clean code (in my opinion), uses unit tests, documents a lot of her code, chooses cool projects, etc. These are features I'm striving to integrate into my own code as I try to transition into programming as a profession instead of just a hobby (and I'm slowly getting there).
I've now taken a sudden interest in either forking or writing one of my own as a learning exercise...and also in reading up more on node.js (which I've been ignoring).
In honor of harthur's efforts, and to take a stand against bullying, I've decided to fork the project. I'm currently porting it to Python!
https://github.com/kennethreitz/replace
I encourage others to do the same in their respective languages :)