l0stman

Well, the wayback machine is your friend. http://web.archive.org/web/20111001124736/http://log.nadim.c...

I was just wondering if you've planned on open-sourcing the code since the very beginning or if the idea came much later. Anyway congrats on launching.

Well, played. For a couple of seconds I was wondering why I had this impression of deja vu then I found this (http://hackerne.ws/item?id=4692794). I burst out laughing.

I had to look up what the `:' operator does and it adds an element to the beginning of the list. So the last line should be

    split (left, right) x = if x<p then (x:left, right), else (left, x:right)

    qsort :: Ord a => [a] -> [a]
    qsort []     = []
    qsort (p:xs) = qsort lesser ++ [p] ++ qsort greater
        where
                lesser = filter (< p) xs
                greater = filter (>= p) xs

I know nothing about Haskell but I don't think this code implements the original quicksort algorithm which sorts the input in-place. Moreover the two-pass of filter over the list and the concatenations cause unnecessary overhead. Thus, even if the sample code is simple and elegant, a real-world implementation would surely be a bit longer than the given example.

Just finished ``On Lisp''. This is an advanced Lisp book and shouldn't be your first one. It assumes some familiarity with the language so you're better of reading ``Practical Common Lisp'' first if you're diving into Lisp for the first time. The book focuses almost exclusively on advanced use of macros and metaprogramming.

Interestingly, I've also read SICP in parallel -- currently reading the last chapter -- and there's some overlap in the topics treated in the two books. I'm also walking through CLRS and TAOCP at the moment, albeit reading the latter with a slower pace compared to the former.

On the non-technical side, I'm reading ``Drawing with the right side of the brain'' by Betty Edwards. I've picked up this book since I've decided to resume drawing one year ago. I highly recommend it if you struggle to draw a realistic rendering of a real life object or a landscape.

Just add the following lines in your .bashrc:

    export GIT_PS1_SHOWDIRTYSTATE=yes
    export GIT_PS1_SHOWUNTRACKEDFILES=yes

Note also that there's a more feature complete version that accomplishes the same tasks in the contrib directory of git (http://git.kernel.org/?p=git/git.git;a=blob;f=contrib/comple...).

It's implemented in bash but I use it with almost no modification with the Korn shell.

Actually I posted this article yesterday http://news.ycombinator.com/item?id=1624402 but I went unnoticed. The author was working on search feature for his Hex editor and tried to outperform GNU grep by using Boyer-Moore. After multiple attempts, he finally discovered the hard way all the tricks used by Mike Heartel to make GNU grep run faster than his implementation.

Thanks for sharing this, it was extremely brave of you to do so. Speaking of one's failures in public requires a high degree of courage. Besides, from reading your other posts on HN, I thought you were what I would generally consider a successful guy. But like you said, there's always a flip side in every situation.