Forget about the page procuring prowess of Google or Technorati’s blog search boom, what the Internet needs most is a way to discover, index, and deliver relevant human conversations to searchers.
Microsoft’s Visual Studio maven Josh Ledgard blogged about what he called “the race for conversation search.” Ledgard knows what conversation search isn’t today:
I’m not talking about a link ranked blog search. I’m not talking about Usenet search. I’m also not asking for a general web search. I’m talking about doing for human conversations across the internet what services like Froogle do for shopping.
Let me listen in to, learn from, and absorb every IRC chat, blog entry, forum post, and newsgroup thread, and wiki article, that’s publicly accessible on the internet and that I participate in privately. How would this be different from a standard web search or blog searches of today?
Conversation search may already exist. But one won’t find it available as a search box on a clean white web page bearing a friendly “submit query” type of button. At least, not publicly.
We have written before about how companies like Factiva and Umbria offer corporate-level solutions to tracking online discussions, and how they are utilized. Those companies, and we’ll likely hear about others, do delve into message boards and blog posts along with whatever websites may mention a corporate client.
They haven’t mentioned IRC, but if there is a way to index those discussions then the companies will likely implement it as well. Factiva described its Reputation Intelligence service as one that does go into message boards and blog posts to find what is being said about a client.
The difference between that and what a typical web search does today is the way the results are conveyed to the client. Visual tools detailed by Factiva go beyond the lengthy lists of search results delivered by engines.
It may be Google that gets to conversation search first. The release of its Trends tool offers a very basic way of comparing two or more search terms to gauge relative interest in them by searchers.
Search terms don’t make a conversation, of course. Assessing that relationship requires changes in search algorithms that, if they exist now, have not been revealed. They do a good job in finding relevance in context. Maybe as Ledgard suggests, some search engine will be able to do even more.
—
Tag:
Add to Del.icio.us | DiggThis | Yahoo! My Web | Furl
Bookmark Murdok: 
David Utter is a staff writer for Murdok covering technology and business.
