If you saw my earlier post about Wolfram Alpha, then you know that semantic search has been on my mind recently. To define it very broadly, a semantic search engine is one that understands human language as it is actually used. Current generation search engines operate by looking for keywords - we've all essentially trained ourselves to ask questions in a way that search engines understand. The challenge with this is that sometimes we're looking for something a bit too fuzzy for keyword-based search.

A good example of this is image searching. Suppose you were looking for a screencap from the scene in The Godfather where Don Corleone collapses in his tomato garden. Google Image Search is tremendously powerful, but it's keyword-based, so searching for this on GIS would entail a search string like "godfather collapse" and then trawling through the results until you find what you're looking for. An ideal semantic search engine could just be asked "Where can I find a screencap from the scene in The Godfather where Don Corleone collapses in his tomato garden?", and it would know precisely what you want. There's a lot of words in that query that are useless to a keyword search engine but that are essential for a human (or a semantic engine) to understand what you're after.

Developing such a thing is no small task, and it's one that humans have been dreaming about for a long time. HAL 9000 from 2001and the computers from Star Trekare semantic AIs: you just talk to them. But this level of sophistication is a long ways away. Even Wolfram Alpha appears to be limiting itself to answering quantitative questions, like "How many oxygen atoms are there in a mole of carbon dioxide?"

Aardvark is a new service that attempts to cut the Gordian knot of the semantic search problem. If the difficulty in creating a semantic search engine is teaching software to communicate like a human, Aardvark suggests that you throw the problem out entirely and just use humans. Aardvark is a social search engine: users are the search engine, supplying answers from their areas of knowledge. It's impressed me so far, not just for how well it works, but for the approach to modern Web development that it represents.

Aardvark integrates with your IM service and with your email. When you've got an open IM session, you can have Aardvark send you a question that someone on the network has asked. If you have a friend on Aardvark that you know might have the answer, you can refer it directly to them. If you ask a question yourself, it will then route to someone on the network who might have the answer, and you'll get a response via the means you asked it in, an IM or email.  It's an elegant system, one that has to be used to be appreciated.

The limitations to Aardvark right now are ones that go hand-in-hand with most cutting edge Internet toys. In my experience most of the questions come from New York and San Francisco, where all of the early adopters are. If your question is about a good place to have lunch near Union Square or about Web development, Aardvark is the search engine for you. The social aspects of the service are also a bit lacking right now. I tried to invite my friend Matt Burton to the service, and Aardvark told me Matt was already on it, but then didn't give me an option to add him as a friend. You also can't see the questions that your friends have answered on their profile pages yet, though if that's a matter of privacy concerns and not a missing feature I've no idea.

For whatever shortcomings it has at the moment, Aardvark is a great idea. If we define "social search engine" broadly enough, we can find some clear predescessors to Aardvark in Ask MetaFilter, Yahoo! Answers, and Fluther. Twitter can also be employed as a social search engine - when I was talking with Matt Burton about Aardvark, he described it as what asking a question on Twitter would be like if he had thousands of followers. Where Aardvark rises above these existing services is the simplicity of its design, and how it integrates so neatly with your IM and email. In this regard, Aardvark is a perfect example of a service that has adapted around what you could call "the Zipcar point".

Zipcar is a service that starts with the idea that there are enough automobiles in the world already. The average American drives to work by himself in a car built for 5 people, then parks that car for 22 hours of the day. There is a huge surplus of car capacity in the world, and Zipcar seeks to make car capacity more efficient. Why own a car that you'll barely use when you can share a car with a group of people, maximizing the use of the car and minimizing the cost of ownership for you? The Web is approaching a similar Zipcar point, but the commodity isn't car capacity, it's attention.

It's 2009. The Web is completely mainstream and people's browsing habits have crystallized to an extent. People are spending a huge amount of time on sites like Gmail and Facebook. Building a new Web property and attempting to build an audience for it means not just attracting users to the site but asking them to displace some of the time they spend on the sites they already visit every day - this is an uphill fight. Aardvark is a service that operationalizes completely around this idea with its design. The bulk of your interactions with Aardvark take place where you already are: your IM and your email. Once you sign up for Aardvark, there's no need for you to ever go to their Website again. A very intelligent design that takes the path of least resistance in the fight for attention.