Yahoo is finally integrating Flickr results for its image search.  My first thought was “yay — its about time they did something to integrate that property into yahoo proper — peanut butter manifesto’s notwithstanding”.  My second thought was, “I wonder how what Flickr is doing could influence how Yahoo sorts its searches?  Better yet, could it actually *improve* a search engine’s results?”

I think the answer is “… maybe!”

Well, this is an idea that Thomas Hawk goes into much greater depth with, but basically Flickr already has algorithms to determine popularity and relevance.  This is governed largely by the way that images are sorted and tagged and viewed by the community at large.  In fact, the community itself determines the relative importance of each photo, which in turn, has an important effect on how Flickr sorts through those photos and the order in which it serves them up if they are searched through.

Could Yahoo! learn a thing or three about this?  Sure.  And Thomas makes some great points about it, but suffice it to say, its a similar, but very different way of understanding importance.  The current model of determining relevance has to do with many different factors, but one of the principle ones involves the pattern of inbound links to a given page for given keywords.  Link structure and the evolution of that over time is what has driven how search engines do what they do.

But for results and data that are derived from social networks, like Flickr in Yahoo — its a bit different story because of the users tag, or even, internally rank (vote?) for specific photos for a given tag.

I think this introduces an entirely different dynamic, because acquiring inbound links in the World Wide Web at large seems like a very different kettle of fish than acquiring the same kind of positive ranking in the factors which matter in a social system. 

But is this an improvement over the way that Search Engines have done things?  I think the answer could be yes.  After all, Matt Cutts at Google has said that Google is trying to dabble with how humans can help the algorithm do what it needs to do.  And what are social networks if not human activity that involves adding, sorting and tagging information in ways that can be infinitely measured and quantified?

Which begs the question, of course:

If social networks can improve the search results of certain topics (and this is a huge presumption), does it make sense for search engines to start acquiring those social networks? 

Not for a content play — but so that it can examine how certain social networks *do* rank data, and so they can have access *to* that, but also, and literally, own the content of those search results?

I have no idea.  But if Google suddenly snaps up Facebook for people search, please reference me in any future posts you might do. ;)