Thinkcorps

Some notes on the recently launched askville. I just signed up (by asking a question).

I’ve been fascinated lately with wikis, and search apps that attempt to condense and organize information in a knowledge management sense—the management of information collected over time, or “expertise” i suppose you could call it. This, I think tails with an interest in data visualization, which is also generally concerned with timeframes.

In any case, it’s interesting how seeing Askville and comparing it to Yahoo Answers helped me understand more about how these services work.

A few thoughts:

My sample question is first answered within a few minutes (I get an email notification). A lot of answers I see (including mine) are copied from other sources—a copyright violation that’s easy to track. I can’t see publishers being psyched about this.

Amazon uses mechanical turk for questions that aren’t getting answers (other people do the search for you). Though I imagine this is a temporary measure until the community grows large enough.

What occurs to me (and maybe it’s the inclusion of mechanical turk that made this obvious to me) is that what drives an answer community (besides entertainment) is not that people are smarter than you, it’s that they can perform a particular topical search better than you.

If people are really using this for answers and not just for fun (unclear at this early stage—yahoo seems very ‘fun’ oriented to me), then this implies that Google returns too many search results. So is QA is a viable alternative/competitor to search? It’s Mechanical Turk turned into a game for getting the easy answers (the assumption being that many things are “easy” with a large enough audience.)

On the other hand, I’ve often wondered what a service like Rollyo could do if it were to use its data to help focus search. With Rollyo’s data prioritizing which sites people thought were good for particular search topics, a search engine could effectively harness that “human power” and get some of its mojo back from the QA communities.

What concerns me about services like Askville, is that by prioritizing the “game” aspect of the community, they hinder the ability to build a detailed knowledge base. The game is completely focused on the “now”: users have to be able to earn points for answering new questions, even if that question has already been answered several times over. And the winning answer in a current questions’ iteration may not be as good as the winning answer of a previous iteration.

Yes, you can search old questions, but historical search is clearly separated from the primary QA process. At no point does Amazon suggest that, “hey, i think we answered that a while ago.” And that’s a problem, in my opinion.

Wikipedia, on the other hand, stresses completeness, where users are focused on filling in the gaps in knowledge and improving accuracy. Of course, Wikipedia also has much higher barriers to entry.

When you make expertise a game, I wonder how you maintain a standard above status quo “popular wisdom”. There’s really no way to gauge authority and no respect of past contributions, pushing knowledge ever further.

Users may accumulate “expertise” points, but those are clearly doled out in a haphazard fashion. I mean, if you don’t know the answer, how do you judge the best one?

It may turn out that QA is 95% entertainment / 5% useful, and I’d guess that Wikipedia would end up the inverse.