Comments on: Build Huffman Compression in Ruby

By: Zack Fonde

Zack Fonde — Tue, 17 Jun 2008 05:50:44 +0000

forgot to mention.... The mainly slow part in my Lisp code is the writing of binary bytes....which I think could be improved if there is some sort of buffering where big chunks are written at once..

Could this be the issue with your code?

Regars

By: Zack Fonde

Zack Fonde — Tue, 17 Jun 2008 05:20:44 +0000

Very interesting, I'm new to Ruby and frankly it amazes me with its elegance... I ran your code with larger text (128 KB) and it seems to go forever, now I had a similar issue when I wrote a huffman coding/enconding in Lisp before, I had the intention of visiting it again to improve on it but that never happened.

I'm wondering if you know where the bottle neck is here in your code, what could we do to improve the efficiency...I'll make this an exercise for me to learn more about Ruby, but would appreciate pointers from you also...

Thank you

By: proj

proj — Tue, 08 Jan 2008 09:22:40 +0000

Akita, thank you. Keep up the great work on your site!

Nitpicker, I didn't implement the weight sorting as you point out. I'll look for an elegant way to approach that tomorrow. Sounds like you already solved that particular issue. Well done.

By: Nitpicker

Nitpicker — Tue, 08 Jan 2008 08:14:41 +0000

I’m pretty sure your implementation is wrong. It should be 312.625 bytes. The reason why it is wrong is because you don’t always combine the nodes with the two lowest frequencies (the interior Array should be ordered).

By: AkitaOnRails

AkitaOnRails — Tue, 08 Jan 2008 04:56:58 +0000

Very interesting! There’s been more than a decade since I first implemented Huffman as a college exercise in Pascal. It is good to see it done in Ruby once again. Programmers have to understand the importance of algorithm study. Kudos for bringing this up.