Hello Martin

Was just reading a NYT piece on how Arthur Schlesinger died yesterday, and i found the newspiece to be a bit to steep since i didn’t know enough about Mr Schlesinger, so an idea poped up.

Someone (You maybe!?) Should build a plugin for all browsers, that analyzes each page you browse (Or a set series of domains of your choosing) and then picks up what say the main 3 keywords for that page is.
For the selection of keywords in most cases a simple Keyword Density checked with stop words etc removed would do just fine.

Data for this page is:
Word - Instances - % of total words on page

1. article 24 2.37%
   
2. click 21 2.07%
   
3. schlesinger 17 1.68%

And for multiple words:

1. more articles 12 1.99%
2. new york 11 1.83%
3. york times 8 1.33%
4. arthur m 6 1%
5. white house 4 0.66%
6. arthur schlesinger 3 0.5%

As you can see both Article and Click could be described as stop words, or “common” words and could easily be removed.

Next step would be to match it up with wikipedia, you could hit Wikipedia with a query straight off, but since the Wikipedia database is available for download i would suggest that someone downloaded it and had a set domain for querying to guarantee fast response.

After this was done a button would appear somewhere in your toolbar or down right hand corner.

Revenue possibilities… well if you wanted to make money you could do a banner layover once the user hit the button to wikipedia… not saying that’s what you should do though.