Kosmix: One step better Search

Posted by Posted on by
0

Kosmix (via) is an example of next generation Search Engine in its nascent stages. It is different from a standard Search Engine in the context that it is Categorized. To start with, they have three categories – Medicine (in Beta), Travel and Politics (both in Alpha). A common clueless web-surfer who uses a Search Engine for generic queries on a regular basis can readily appreciate the difference. Since the categories are limited, queries have to be, for the time-being, specific to these topics.

For instance, there is a huge difference between searching for a Medical condition in Kosmix and in a standard Search Engine. I entered “Avian Flu” as a search query. The results are presented in an extremely ordered and sorted out manner. For each search term entered, there are sub-categories viz., Basic Information, Definition, Treatments, Symptoms, etc. But Kosmix search goes a step higher by linking Blogs and Forums that provide information about a condition. However, the best part about Kosmix Medical search is the links it provides to authoritative websites that address the problem of specific patients such as children or women.

As with standard Search Engines, search can be personalized. The Travel category is neat, with links to Travelogues and user reviews. Kosmix definitely has a long way to go before it becomes a next generation Search Engine. But this is a welcome change from using quotes, special characters and advanced search filters (and not to mention a great deal of effort involved in thinking about the right set of keywords) to find the information required. It seems that Kosmix was launched by two Stanford graduates. What is it with Stanford Graduates and Search Engines? ;)

Social Search?

Posted by Posted on by
0

“Social” is the buzzword these days. Social Bookmarking, Social Networking, Social News-sites, etc. And so Microsoft is going the Social Way – Social MSN Search. To quote:

“The feature will let users direct questions to a specific universe, such as a group of friends, rather than to get automated lists of results from a generic search engine.”

The article reports that Microsoft will tie up with Eurekster which is community powered Search Engine. Sounds a nice thing to do, especially since a generic search most certainly returns irrelevant results; also the likes of Friendster or Orkut are potential sources of lot of information that’s poorly organized. Sometime back, Seth Godin illustrated this idea in his blogpost Tea. The method wasn’t right but the idea is. Hand-built search engines seem to be doing real good job when it comes to searching for generic search terms and recommendations for them. Digg is a fine example for articles recommended by other readers.

There are several Community powered Search Engines – full list can be found here (under the Search 2.0 Category). Most of them enable “rating” a search result. And the relevance of a search query improves as more people use it. So what is special about this Social Search? In what way is it different from any of the Search 2.0 sites?

Consider an option of buying a camera A or B at a place C or at a place D. There are several shopping websites that give you a price list. So enter the name of the camera you wish to buy, and you shall be able see the results and product reviews by people who bought it. Picture this against posting a question: “Should I buy A or B?” to a community dedicated to Digital Photography. The answer will be personalized and answer your question directly unlike the fifty plus search results.

Another example. This time personal. Recently there was a movie released by the name “The Island” which is a dystopian story. Aldous Huxley wrote a book “Island” which is of a similar theme. I wanted to know if there was a connection. Google and Amazon didn’t provide with satisfactory answers. Orkut’s Aldous Huxley community did. And I didn’t even have to scroll my way through several responses.

Many people turn into search engines for answers to their questions. They do not want search results for a key word. This is where I think community search will make all the difference.

Overbrain: Or can the web be “intelligent”?

Posted by Posted on by
0

I happened to read an editorial in Febrauray issue of Digit magazine the crux of which was “how can the web be intelligent”. And more importantly, the article started of with a familiar example – “Customers who bought this also bought” section in Amazon.com. When I first chanced upon Amazon a couple of years ago, I was drawn to this feature like how . It was a great place for book recommendations, and in many cases it worked out fine for me.

Back then the web wasn’t run by people, like it’s today. There weren’t any web-apps developer or a personal database which is, well, more personal than a website like Amazon. Then slowly blogs and tagging evolved. And to the relief of a lot of people, web-apps can be created in no time. It seems almost true that the internet contains information about anything under the stars.
Which presents the most important problem of Information Overload. A couple of months back, I entered a search query concerning something and got a zilch response. But a few days back, I accidentally stumbled upon what I was looking for. My joy knew no bounds on hitting the jackpot accidentally. But it also made me crib for it had been there all along and I’d wasted a good three months in searching. Not that my search skills are that bad, (even if I enter the same query now, I get no results). They are not what they have to be.

Let’s assume this hypothetical case. I’m looking for information on a particular group who will be performing at the local theatre. I enter their name, I get no results. Not even “are you looking for” suggestion. I tend to feel that their company isn’t adjusting too well to the internet boom. Later on, I almost accidentally stumble upon the group’s blog, and going through their archives, I come to know that they’ve indeed been existing in the virtual world for a long time, that my search query should have landed me here by all means. What happened to me was nearly similar to this situation.

This calls for a need to orgainze content better. The article from “Digit” calls for a “People who visited this website also visited the following other website(s)” method. This is okay as a next step to better search. Like Del.ico.us. But given the number of websites and tag clouds, it is testing the limits of exhaustion on the readers part. Whether an algorithm can handle this or human intervention is required at all stages, only the researches can answer. As a person who searches the web for information almost everyday I’ll get a sense of absolute satisfaction only when I enter a search query like, “Is tomorrow a Friday”, and get a reply, “Yes, TGIF”.

Edited to Add:

(1) In fact I was thinking of Thandora when I wrote this.

(2) TGIF means “Thank Goodness it’s Friday”, which, for most people, is a relief.