While I'm not ready to put any code online, I thought I would make a brief note about the latest phase in the development of my online note taking application.

At some point in the last week, I decided it would be useful to give my software the ability to algorithmically generate short summaries of long notes. From the point of view of the user interface, it's quite a bit easier to browse notes and find something when the software doesn't display too little information — just a title or some statistics — or too much information, such as the whole ten paragraph note. (Think how much scrolling users would have to do!)

I've spent the last two days working on the algorithm, and I'm really quite pleased with how well it works, especially since it's a relatively simple algorithm, of less than 300 lines of code.

For instance, I just ran an article from CNN about Sarah Palin's recent decision to step down as Governor of Alaska through the algorithm. The original article is a little more than 5800 characters. My algorithm produced this 54 character summary:

It has a number of flaws, like the stray quote marks, incorrect punctuation of questions and a pronoun as the first word, but it's not half bad either.