Tuesday, December 06, 2005

what do I know?

Hmm, you all don't seen very interested in my non-tech writings, no one ever seems to comment on them. Sadness, I really do think (hope) my musings as a person is much more interesting than my ramblings as a PM or armchair CEO. Well, so be it...I shall indulge you w/ some more technobabble then.

No, really. Believe it or not I spent a lot of time thinking and writing up that post on relationships - much more than whenever I spout off on tech!

********
I was complaining about my overflooding feed list to Huat again the other day, about how I never have enough time to catch up on all my blog reading. Plus I keep subscribing to new blogs as existing blogs keep linking outside and the network keeps expanding. On the other hand, I've been able to keep my inbox at zero for pretty much the last 4 months now, clearing things at the end of the day and it's worked very well. I wonder if there are analogies of email management that can be applied to feeds?

Could one setup some rules to manage how items are prioritized and read? Obviously the difficulties are there are no "to recipient" rules (obviously) and the "by sender" rule is less useful as authors can write about anything under the sun. And rather than grouping into a tree by author, what if you re-sorted items by topic or relevance? Can some kind of contextual analysis be applied to the content, where the machine attempts to parse the feed content and decipher what the feed is about, then sort it for you in terms of importance, or perahps group by topic? Can you apply PageRank to blogs (actually does google or technorati do this already? I may be woefully outdated) but internally, so rather than having to query an external index using some keywords, what if you flipped the index inside out so that the keywords or nexi (plural of nexus?) are exposed first - basically clustering applied to feed items. Again, what I find so valuable about blogging is its conversational and never-ending expansive nature, as blogs keep linking out ad infinitum. What if you not only clustered your subscribed feeds, but also crawled all their outgoing links, and links to those links, and so forth, and clustered that?

I guess what I'm really looking for, in my search for the next big idea, is a tool to visually organize and present this vast web of discussions out there, so that I can easily filter out the nuggets of gold among all the pebbles, and draw my own connections and conclusions. Actually, I bet there is probably already a reader out there like that, I just don't know about it yet...