).Note that I've not modified any text so may seem a bit odd here and there. All are welcome to comment and share your thoughts :)
There has been a lot of talk about faceted search and this is high time to give all the enthusiastic readers a hint of where we are heading with this. The research crowd will immediately know (or at least remember of vaguely hearing about) Faceted Search but its too much to expect from the common man :) So here goes - a quick take on faceted search and what it brings to us.
Imagine that you are looking for something to do over the weekend and decide that you want to go for a movie. There are a lot of movies and searching through the movies and deciding what you would enjoy is not so easy. If we put on our analytical hat, you'll see that we can look at this problem of selecting a movie in several angles. First lets decide what makes you like a movie. some people (like me :)) love adventure/action/thriller type movies. There are others who enjoy other genres (say horror, romance, comedy etc). Some like movies because of the cast. Some like movies by a certain director. What is important to remember in this is that all of this happens over the same data. You can say 'Bourne Ultimatum' is a action/thriller movie as well as a Matt Damon movie, and each point of view would be interesting to a different party. Each of these so called angles is a facet.
Here is one of my own examples. Think of a crystal cube (No not a crystal ball - a crystal cube. Similar to the ones that a fortune teller would use :)) This crystal cube has a some blob inside it and each face of it is of a different color. When you look through the red face - you see the blob as a sphere. When you look through the blue face, you see the blob as a cone , similarly rest of the faces giving a different perspective of the blob inside. Now let us rewind back to our movie example and think of the blob as movies. One face shows you the movies categorized by genres. The other face shows you the movies by the cast and so on. This is exactly what faceted search is. It gives you the ability to look at a data set through a facet..Forgive me for my crude drawings but the following image should drive the point home.
Why is this important or interesting at all ? Going back to our movie example, I would immediately go for the genres and look under action. namelessnerd might (?) go by the director (he is a big fan of Stanley Kubrick). But this is not as simple as filtering using the director column. Let us say you want to see the user ratings of the movies also. To make sense of it when you look through the genre facet you should see ratings of people who liked the movie because of the genre. Same applies to other facets. This makes decision making so easy and you can quickly get to what you want. Another interesting thing you can do with facets is to cascade these facets. First look for the genre and then the cast and so on.
Ok - Now to the important part. We want to apply the same concept to searching for Web API's. Any geek (or anyone who would have played with any of the Web API's out there) would know how hard it is to mix and match different API's (of course the situation has improved a lot. But still things are not perfect or seamless). Given that it is inevitable to avoid a bit of manual labor (:)) we think that a faceted search for API's would make the life a lot easier for people. For example you would be able to say "show me all the services that support RESTful invocation" or "Show me all APIs that provide image manipulation". Things don't stop there and we have a whole line of neat tricks up in our sleeves ;) Hopefully this gave all these interested people a glimpse of where we are heading.
