Devangshu Datta: The hunt for a perfect search engine

TECHNO BEAT

In any lexicon of 21st century usage, "google" has the same status as "xerox" in a previous era. There are hundreds of competing photocopiers available.

But xerox's pioneering status is enshrined by the brand name's synonymity with the act of photocopying. There are many search engines but one "googles" when one is looking for online information.

Google remains the leader in search engine technology. But its position is being challenged by others who have created alternate versions of Google's sort-and-search methods.

Google's technology revolves around the patented PageRank, which weighs webpages on the basis of the number of links, the traffic, keywords, and other factors.

The massive number of pages it has indexed and cached makes Google the engine of choice. For 10-year-olds cogging homework, for scientists looking for papers, for perverts hunting for pregnant grannies, Google is the place to be.

If you're looking for something, it's likely to be there in the cache and by inputting the right keywords, you will find it without too much trouble.

But page and link analysis as such cannot be patented although you can patent a given algorithm. Microsoft would like to change Google's paradigm of market dominance in 2005 by creating an even better search engine.

Google's revenues ($805 million in the July-September 2004 quarter with $11 million profits) are derived from the focused advertising drawn by user-searches. Microsoft would like a slice of that market.

The earlier MSN search engine was driven by Yahoo! and apart from logos and ad-content, you got the same results on MSN and Yahoo! if you made the same searches. The new beta search engine at http:// beta.search.msn.com/ offers substantially different results.

Some website-owners have claimed that MSN indexes and lists new sites quicker than Google. This is perhaps because MSN has "only" 5 billion indexed webpages to Google's 8 billion and indexing proceeds quicker with a smaller database.

It is also being claimed that in a given webpage, MSN's proprietary technology places less reliance on back-links and more on relevant keywords. Weightage to more links, as in the Google method, is built on the assumption that a highly-linked page has more useful information.

More reliance on keywords suggests that a given page may be more relevant to a given search. Both weighing methods can be manipulated by a clever webmaster.

Apart from this, MSN does offer dual language search preferences, which might be useful in certain circumstances. (Google's default setting displays results in all available languages or in one specific language whereas MSN can be set to display in, say, German and English.)

MSN also offers graphic settings that can display "approximate" matches by degree and the popularity of a given webpage. MSN also has a music section where music can be searched for and downloaded at $0.99 per tune. Plus, its online calculator can handle simple algebra.

The last might wean mathematically-challenged students off Google but will it change market equations? MSN currently doesn't compare with Google in terms of spelling corrections (where the user misspells a search word and the site makes educated guesses). Nor does it compare in terms of local listings and the range of language options.

By January, when the MSN search engine will be formally launched, MSN might have got further on these fronts. Google has already fought back with a new toolbar, the Gmail service and Desktop Search, and it will presumably produce more features. The competition will be good for every user and online advertiser.

In other areas also, competition has forced Microsoft to adopt new strategies. In 2005, it will start offering stripped-down versions of Windows XP at a discounted price for specific markets, including India.

This will have a local-language option and launch at around $50 equivalent, which is a third the price of the current Professional Edition. Competition is always good for the consumer even if, in this instance, it comes from the free Linux and pirated XP editions!