needle in a haystack

i tried out wolframalpha today. went there from a slashdot article. the article talked about how, for the first time, some software is claiming copyright on its output. anyway, the search engine or 'the computational knowledge engine' as its creators call it, produces some truly amazing results. while there are already lots of semantic search engines on the net, the little that i dribbled with wolframalpha, made me like the way its results are displayed and also the depth of information provided. i don't know whether there is a law dictating that depth and breadth of information are inversely proportional but it's true that this engine will not turn out results for many queries. so, for example, red hot chilli peppers will give you nutritional information of chillies and peppers. but try searching for an english word or 'who killed kennedy' or 'india' or some stock and the differences will stand out.

however, i'm not sure how all this 'making sense' of queries will make things better. true, it brings you relevant data quicker and in a more compact form - data for which you might have had to visit 2-3 pages. but the presence of information repositories like wikipedia means that, in most cases, if your query makes some sense, you'll get all the required info from wikipedia and links to more information sources. but most of the searches that people do are for queries that don't exactly 'make sense' in the first attempt - these are the queries that matter because information on others is amply available. try searching for 'rural income trends in india' and i'm sure it will be a google or yahoo that will give you your desired result.

however, semantic search is touted to be the next big thing in search. and i suppose i'm not aware of all the uses it can be put to. let's see..