[Standards] Fun and games with Huffman

Klaus Hartke klaus.hartke at googlemail.com
Tue Aug 4 10:23:14 UTC 2009

2009/8/4 Dave Cridland <dave at cridland.net>:
> So, I've been playing with compression algorithms, by way of fun and games.

Interesting. I'm toying around with the binary encoding used in
Microsoft's WCF in my implementation of XMPP end-to-end messaging. It
encodes the XML basically by representing the sequence of parse events
in a binary format and performs reportedly much faster and uses less
space than the alternative XML encoding. I'd be interested to see
what'd happen if a compression algorithm was applied on top of this.

I'm looking into the WCF binary encoding mainly because it looks much
simpler than EXI. For example, <doc></doc> is represented as 40 03 64
6F 63 01 where 40=start element, 03=3 characters, 64 6F 63="doc",
01=end element. There's has a mechanism for abbreviating frequently
used names and prefixes, but that are all optimizations already. More
examples: http://msdn.microsoft.com/en-us/library/cc219284(PROT.10,loband).aspx


More information about the Standards mailing list