Your RSS reader OPML file can reveal a lot about you, just read Alex’s post (which also references Lisa’s post) on how you could in turn share this data to a service so they can recommend stuff, just like Ultragleeper can recommend feeds.
Your OPML file not only reveals your interests but your behaviour or degree of interest (attention data).
Some RSS readers can automatically delete stuff or run reports on feeds that haven’t had new posts, more interesting is that it can also tell your most visited and viewed, and least visited and viewed feeds.
But this type of information is kind of ambiguous at revealing your degree of interest as you may click and view stuff, to realise you don’t like it, so the RSS reader doesn’t know the consequences of reading an item, ie. whether you liked it or not, also sometimes you click through because of the difference between summary and full-text feeds
…flagging an item is more definite.
Along these lines I did a post agreeing that explicit voting or rating a post is more reflective of your real interest, this is also a better method in ranking your feeds that don’t post often.
Only drawback is what if I couldn’t be bothered voting for a post if I have already read a very similar post previously, the second feed wouldn’t seem as important to your RSS reader statistics, but you still like reading this feed…I think that’s why voting and visits/click will work hand in hand.
If we have already read a similar post we probably couldn’t be bothered voting for a current post, but I suppose you have already read a similar post because it is ranked higher in your RSS reader list.
More on your OPML at attention data at DLTQ:
“OPML has so much potential. If I spend time on my OPML file, I will eventually build a multi-layered outline of my entire public life as well as parts of my personal life. For Amazon, this file would have some value. It would also have value for potential customers, employers, or partners in projects. With the growth of OPML browsers, we will be able to surf such accounts of people’s lives.”
Here is almost a real life scenario of OPML attention data, the items in this sort of OPML file are just text, but the idea is that a service could absorb this OPML file and recommend stuff.