Mike Davidson’s article Newsmap as a Model for Smart Aggregation provides an insightful view of the future of information retrieval, presentation and use. I readily admit to being an information junkie – I’m subscribed to several lists (evolt, WebDesign-L, FreeCycle etc.), forums (eZ publish, KnifeForums (work), UsualSuspects (work) etc.), 72 RSS feeds (actually 73, now that I am aware of Mike’s site), several sites that are not currently RSS-enabled, a couple of IRC channels, and a select group of IM contacts. As noted in the article, it isn’t easy to keep up with the steady flow of information, and at times I am required to shut down some of the sources to ensure I can work without distraction.
Most days I rely on a few configuration controls to keep from being overwhelmed by information. Beyond the standard spam controls (SpamAssassin with Thunderbird‘s junk filtering), I use message filtering and labeling to sort incoming mail to folders, delineated by subject or sender, allowing me to focus on important e-mail first. I have set up my RSS aggregator (SharpReader) to refresh a few select feeds every two hours, while the vast majority are refreshed every four or eight hours. Some days I find that the information provided in the less-often refreshed feeds is much more interesting/important than in those feeds I have deemed worth refreshing every two hours. While I can mark myself away or shut off my IM client (Trillian) if I need to focus for a while, there are those occasional periods where I need brief moments of intense concentration in an otherwise relaxed day, only to receive a pop-up message window, distracting me from the task at hand. Pretty damned frustrating to say the least.
Obviously, I am a prime customer for a personal contact/information agent. If I had a tool to control the flow of all of these streams I would be much more efficient, without losing the valuable connections required by my job in a fast-moving industry. I would actually add a couple more features to Mike’s idea.
I have willingly ranked 250+ items on Amazon, in the expectation that I, as a customer will receive better recommendations when viewing their site. I expect Amazon to try to sell me items, they might as well be items that I find interesting, right? Right. Why not tie this type of information/preference into the system as well. If I have assigned a high rank to a book, CD, or DVD on a retail Web site, it would make sense that I might be interested in a notification of a new album by the same artist, book by the same author, or the sequel to the movie. Conversely, if I have given low ranks to an artist or author, then the system could imply that I really don’t want to see news items about them.
If I am discussing Margarita recipes with Leesa via IM, it would be great to have the electronic agent run a search for Margarita recipes and notify me of their results. This would be a very complicated task, as it asks a lot of the agent: which words/phrases should trigger a search, and when is it appropriate to notify me of the search results. For that matter, how it chooses to notify me is as important. Perhaps it would be best to just make an entry for me to view when it is convenient for me.
There are more attributes that I would want to see in an application of this form, but I will save those thoughts for a future post.