Where does Wolfram|Alpha fit in the space occupied by search engines like Google, that have re-defined 'search' being used extensively by anybody and everybody?
First, it's important to understand that Wolfram|Alpha isn't a search engine, but rather a computational knowledge engine. A search engine compares input text against its index of web pages, and returns a list of matching links for you to follow. But Wolfram|Alpha interprets your input and computes a specific answer and analysis on the fly, using a massive internal knowledge base that covers virtually every domain of human knowledge. This means Wolfram|Alpha can not only answer simple search-type queries like "information about China," but also complex, computational queries like "weather in Kolkata two weeks ago," "current distance from the Earth to the Moon," or even "who was President of India when Shahrukh Khan was born?"
While Google is a search engine designed for users at large, do you think Wolfram|Alpha will have elements to meet the needs of mass users?
Part of Wolfram|Alpha's mission is to "democratize data" - to make expert-level knowledge available to everyone, regardless of education or profession. We've made it possible for everyone to get quick, accurate answers to factual questions, without having to search through hundreds of web links. New features in Wolfram|Alpha Pro even allow users to feed their own datasets into Wolfram|Alpha for an automatic, instant analysis. And we're constantly expanding into new areas of knowledge, particularly in more consumer-friendly areas like sports, music, movies, and more. Much of that content is specific to the United States right now, but we're also adding more localized data for users around the world.
While part of your mission is to "democratize data", can we expect something for everybody in near future?
If you look at www.wolframalpha. com/examples, you shall get a feel for all areas of coverage we already have. There's already plenty there to catch the interest of just about anyone -- from movies to weather conditions to stars overhead to measurement conversions.
What semantic web technologies do you use to make the engine smarter than traditional search engines?
Wolfram|Alpha is not searching the Semantic Web per se. It takes search queries and maps them to an exact semantic understanding of the query, which is then processed against its curated knowledge base. The main technology used is Mathematica whose language is used to describe the semantic queries, and Mathematica technology is used to build up the natural language parser, the data curation pipeline and perform the data processing, computation and visualization.
Where does the raw data come from? How do you deal with problems of inconsistency, since data comes from myriad sources?
Wolfram|Alpha's knowledge base draws on thousands of different public and private datasets, all of them carefully reviewed and curated by experts in each domain. You're right that there are often inconsistencies and other problems with data from different sources. In some cases, including certain scientific domains, there's a reasonable range of values for a particular answer, so we'll represent that data as an interval or show an appropriate margin of error. In some cases, we're able to use the power of Mathematica to synthesize a single answer from multiple sources. And in many cases, we find that inconsistencies are simply due to errors in source data; our data analysis and curation tools regularly uncover bugs in datasets from well-known data providers - something that a regular search engine simply isn't designed to do, but that results in better-quality data for everyone.
We see a slew of Wolfram-based mobile apps available for iPhone/iPad and Android platforms. Do you have apps for SMEs?
We are starting to develop along these lines. In addition to Course Assistant apps, we have also created a few Professional Assistant apps and personal apps. Right now we have Professional Assistant apps for network admins and lawyers and very handy personal apps for travel and finance.
How do you see search engines being used five years down the line? Do you think, knowledge-based engines shall get prominence in near future?
As the number of web pages increase over time, the traditional search engine approach of finding pages that match keywords in the user query becomes less and less meaningful. What does one do with a million plus results? Already, you can see how search engines like Google and Bing are trying to add additional features to prioritize and rank pages better to give answers to what the user actually wants to know. Users want answers, and that's why we believe knowledge engines will eventually be used more and more commonly.