[markessien]

What I'm thinking of here is a search engine that returns a single result per item, and then if the result answers your question, it ranks higher in the search result, meaning it answers other peoples questions too. You can click a button to get rid of the current result and view the next. This way, the search engine always gives the best possible reply.

[gojomo]

Was thinking of something like this, also.

I could see it working, but mainly for queries that have one clearly-best result, or where the 'snippet' itself answers the query. So perhaps for 'answers' search more than general web search.

The implicit user feedback created would be great... but uses might be frustrated by not being able to scan many results quickly. (I always set Google to return 100 results...)

[markessien]

At the start, it would suck, but if the search engine reached 98% accuracy rate, you don't need more results. The engine could even learn when results are in order - for example, asking "where is mongolia" will give you a single answer, but "tv review sites" will give you a list of search results.

[gojomo]

So bootstrapping is a challenge. Will users wait out the 'training' period?

(Or similarly, if a forward-thinking group is willing to help train, are their choices representative enough of what the impatient masses eventually want?)

I think it's a promising area for experimentation.

[markessien]

I am going to model this mathematically soon, but with a very small sample size, you can quickly reach a pretty accurate result. And in general, most results are pretty clear - number of states in the u.s will always be the same, no matter how forward thinking you are.

Such an engine has to start small and be trained for a few months, otherwise users will think it sucks. As it grows, it adapts.

I bet one could actually do this very easily using yahoo BOSS. Want to take a stab at it? If anyone wanted to work on this, I'd help, so long it was written in python.

[litewulf]

Forgive me if I misunderstand, but couldn't you just look at what people click on and assume they had clicked next for all the results before it?

[markessien]

No, the point is that you have a text box that always accurately answers your query. For example, you type in "hangover cure" and it gives you a hangover cure. You type in "what is the capital of Benin" and it gives you the answer. You type in "Microsoft" and it shows you the microsoft website.

The search engine learns because when the answer is wrong, people select the "next" button. If you just return a list of search results, people will often click the first link because it's first.

[mkuhn]

this somewhat sounds like the search-wiki extension Google has. you can only see it when you are logged in but you can influence the ordering of results. it's not one result but you can manipulate them and AFAIK Google mentioned that they want to use that data sometime in the future.