Slash, Tag

Those who follow global technology developments may have heard about Rich Skrenta. One of his earliest achievements — while still in school — was to write what probably was the first self-propagating computer virus. He is better remembered for founding Topix, a news aggregator site that displays news depending on where you are. Topix made Skrenta rich, as he and his fellow founders sold 75 per cent of the stake to some media firms. Now, he is back with a new venture: a search engine with a mind. Or, to be precise, with a multitude of minds.

Skrenta started work on search engine company Blekko in 2007, opened it up for a few people in July this year, and launched it officially early this week. Blekko provides search results that appear to have been edited. In fact, they have been edited to weed out useless websites. Blekko employs people to do this editing, and they comb out websites with content designed to be on top of a Google search. You can do the editing yourself, if you wish, by creating what Blekko calls slashtags. Here is where Blekko differs from Google.

A slashtag is a search query — a word usually — punctuated by a slash with another query. The second query word turns the search from a horizontal one to a vertical one. It turns a general query into a specialised one. For example, if you put the word health after the slash, you get only results from trusted health websites. A random test shows Google returns better results, but Blekko often blocks out content that is useless. Of course, Blekko does not give results biased towards India, as Google does if you search from India. Neither is what Blekko does completely original because other sites, including Google, have tried out some of its concepts in various ways before (see ‘Predecessors').

Skrenta had said that he started the company primarily to shun the hundreds of thousands of useless sites that turn up periodically in a typical — Google — search. He specifically means the content farms, websites that employ non-specialists to churn out millions of pages of text. Blekko claims to be able to prefer content developed by specialists over those in content farms. It has developed a list of seven categories that people often use, and it uses slashtags in these categories even if you do not. Search engines such as these require time to improve their method. Let us wait for a year or two to see where Blekko goes.

THE PREDECESSORS

Powerset: Was building a natural language processing engine when it was acquired by Microsoft in 2008 Cuil: Set up in 2008, it tried to organise the search results. Was shut down two months ago Hakia: A search engine that uses a ‘semantic rank algorithm', and tries to rank pages according to meaning Mahalo: A search engine powered by humans and not algorithms, Mahalo has profiles about commonly used queries Exalead: Provides search platforms and search-based applications for businesses. Uses semantic technologies