A programmer's basic instinct is to automate any manual task you find yourself doing repeatedly. That's why I'm amazed we haven't built better solutions for finding people online. Most people go through the following steps when they want to know more about someone they've just met:
1 – Type a name into Google, LinkedIn or Facebook and see what public profiles appear.
2 – Figure out which profiles are for right person based on what else you know about them, either a rough location, their job, friends you have in common or other sites they list.
As more and more information is published about individual users step two gets easier, because you can cross-check across multiple accounts. Maybe my LinkedIn profile doesn't give enough details to be sure that I'm the Pete Warden you met, but it links to my Twitter account where I'm rambling about my upcoming UK visit, and that fits with the funny accent you remember.
What's missing is a good set of tools to assist the second step. It's silly to have people wasting time doing this sort of detective work manually, when some simple automation would speed up the whole process. The data on Twitter, LinkedIn and other public profiles has some structure, it just requires some smarter indexing on the search engine side to make use of it. My Twitter profile lists data in hCard format so it's easy to figure out that http://twitter.com/petewarden is about a person called "Pete Warden" based in Boulder, CO. My LinkedIn profile also uses hCard and describes a person called "Pete Warden" in the Greater Denver Area. Why not make a wild guess and present all the profiles that are close matches like that together in the search results? Sure, the grouping will be wrong sometimes, but most of the time it will cut out a lot of messing around on the user's part to do the same process manually.
Google's Profiles would be great holders for that sort of information, but they require users to fill out yet another set of forms. Sites like 123people.com try to automate the whole process, but frankly don't do a good job and are packed with off-putting ads.
It's the spread of services like Gist, Xobni and Rapportive that gives me hope that change is on the horizon. Data flows into them from either their own customer base or providers like Rapleaf, and they're starting to build unified pictures of people online. Just like a phone book in the old days, you should be able to enter someone's name and get whatever information they've chosen to publish about themselves.