« Being bookish: events tonight   |   Main   |   Jeff Hawkins at ETL today »

Capturing the full view from microcontent

After a long and torturous session, my subscriptions are finally back in order. In a dramatic instance proving subject-based taxonomies don’t always work, I’ve smushed my subs into a few broad groups. The grouping’s based on a mix of knowing people, interest in content, posting frequency, and - perhaps strangely - ratio of “new” to rehash.

I’ve got a lot of feeds, but wasn’t trawling them effectively - but when I do, I’m struck by the ‘yada-yada’ effect of many blogs. And that’s my posting style - if I can get to something news-ish immediately (what Jeff noted here) I’ll post on it, but as it decays, I care less. It comes down to the fact that for me there’s a short window where you can legitimately repost facts, say an acquisiton or rumor, without providing analysis. After that I expect to be told something new. Which isn’t really fair, since the order in which I cruise a blog has nothing to do with when it was written.

At long last I powered through all my feeds yesterday, and have such a sense of Greek chorus. We’re all singing the same song - which would imply I need fewer feeds. But that’s not really true.

This gets back to the end-all-and-be-all that I wish someone would come up with - a conversation tracker. I’ve talked with the guys at Technorati about this, and the complexity is insurmountable they say. I bet that there are folks out there with a more pragmatic, just get it working ethos that can do it. This isn’t a criticism, Technorati works on a theory of precisely mapping the known and countable. This project would be a “good enough” tool, and might therefore be a good candidate for a Google beta (though it would be very dangerous for Technorati since it would be the idea pretex for encroachment on their market).

You’d need to collect all the feeds from all the blogs out there and do processing and mapping to very imprecisely map out the conversation. Me, I’d like to see it by date, and watch the flow of the conversation as the snowballs move up and down the hill. Others might want to watch it by authority. Whatever works for you - it should just be a sort on the data. When we talk about AttentionXML - the problem that solves for me is to help surface what I don’t already know or have access to. In all of these discussions, adding a time-based linking trail would help map out discussions, simulataneously tracking their linear progresssion as well as the collective mind map to capture how ideas morph and change, merging on later with other themes.

And this is the same kind of text processing and intelligence that Paul Kedrosky was expecting from PatentMojo. And when it comes to search, getting beyond keywords to context requires much more heavy lifting.

What I wonder based on this, and talking with the Technorati guys over the months is the degree to which user bitching in the “you can’t find my important post” works to really constrain the innovation that a company is comfortable with. Google’s clearly ok with alienating whole tranches of society in their experiements-cum-product-offerings, but they run their business to be isolated from this. The degree to which Technorati or Six Apart, as members of the blogging community, seem to have their moves questioned and second-guessed by the ecosystem is a striking counterweight to their ability to make the kind of leaps necessary to innovation.