Idea: Dynamic compression

[i-rme]

Currently itty-bitty uses lzma for compression, this is ok for general content but I believe that trying language-optimized dictionary compression methods could be interesting.

Here is an idea for a possible approach:

• Implement several compression algorithms (lzma, html tags, english dictionary, spanish, both, etc)
• On link creation, the edit page should try all the compression methods to decide which one of them is more suitable for the text requested.
• The first character of the URL would indicate the compression algorithm used.

This way we could have shorter urls if the user writes an article using common words in one or two languages.

I don't know how much this approach could help, but here is the idea if someone wants to do the math.