TT Epaper
The Telegraph
 
 
IN TODAY'S PAPER
WEEKLY FEATURES
CITIES AND REGIONS
ARCHIVES
Since 1st March, 1999
 
THE TELEGRAPH
 
CIMA Gallary
 
Email This Page
A question of answers

A cartoon in a US newspaper summed up how our lives have been changed by knowledge-at-our-fingertips technology. A man gazes thoughtfully into space and says: “I’ve been wondering about the meaning of life.” “Ask Google,” comes the instant reply from his wife.

But ever since Google came into our lives there have been rivals trying to outsmart it with better answers to our millions of multitudinous queries. Now a new smartie pants is swaggering into cyberspace with promises of being quicker on the draw and faster with the answers.

It’s called Wolfram Alpha and it’s the brainchild of Stephen Wolfram, the creator of what’s called the Mathematica symbolic computation system and author of a pathbreaking book called A New Kind of Science.

Wolfram has a Google-sized vision about the way we use the Internet to look for data. His new system is already making huge waves and creating an ever growing buzz on the Internet — even though it hasn’t even been started yet. It’s scheduled to launch on Monday (at the time of writing this article).

Its inventor calls Wolfram Alpha a “computational knowledge engine”. That’s quite a mouthful, right? What’s it all about? Imagine entering a query into your favourite search engine and, instead of getting a list of links in return, the search engine actually answered your question. That’s Alpha for you.

To call it a traditional search engine would be a misnomer. It’s at best a “fact search engine” or maybe an “answer search engine,” — it is designed to provide you with direct answers, rather than point you at pages that in turn may hold those answers, as in the case of Google.

Let’s take a step backward to explain how Alpha differs. Over the past four years, a huge team at Wolfram Research has worked in secret and licenced — or created — a whole library of databases and manipulated them so the information is pliable. To date, they’ve included terabytes of data from Wikipedia, the US Census, and “about nine-tenths of what you’d see on the main shelves of a reference library”.

But let’s quickly kill another misconception here. It’s not an encyclopaedia either. In fact, most of the information on any topic in Alpha is rudimentary and derivative. It’s great with recorded numerical facts though — enter a country’s name and it will give you its GDP, GNP, size, population, but not much about its history. Think of it more as a reference almanac that’s been put on computational steroids, and you’ll begin to scratch the surface of what Alpha can do.

What this means is that you can look up the population of New Delhi, the height of the Eiffel Tower or the temperature in Calcutta today. But the real power of Alpha lies in its ability to take data that can be numerically represented, and calculate new relationships out of that data.

For instance, which has the larger population — New Delhi or Calcutta, and by how much? Fancy comparative population growth charts over time? Check! Looking to settle Guinness Book-style disputes — oldest man, highest mountain in China? Check!

What is the height of the Empire State Building divided by the length of the Golden Gate Bridge? Crazy, but Alpha spits out an answer in a blink. Too simplistic, you say? How about we juxtapose data, to pull in a totally different set of facts and form an entirely new correlation? Think about it — you compare population trends to the amount of fish consumed and correlate it to mortality rate. Wow.

For anyone with any amount of interest in maths, Alpha really blazes a trail. Why wouldn’t it — it’s got the genius of the Mathematica software behind it, so in a sense, it’s basically a Mathematica front-end capable of deciphering natural language queries.

Try asking Alpha the result of sin x/ cos y, and it spits out the result, complete with a contour plot, a 3D plot, derivatives and roots and a whole lot of mathematics that I forgot way back in high school. Wish I had Alpha at hand when I was trying to figure out the complexities of integration and differentiation!

And it’s not just complex college maths that Alpha can attack. If you ask it everyday calculations such as compound interest, EMI calculation or body mass index, you’ll also get a new page with a form which has fields for principal, return on interest, time etc.

Another interesting application was a query with a very sophisticated result — say, “uncle’s uncle’s brother’s son.” Now if you type that into Google, the result will be a useless list of sites that don’t even answer this specific question, but Alpha actually returns an interactive genealogic tree with additional information, including data about the ‘blood relationship fraction,’ for example (3.125 per cent in this case). It’s impressive, any which way you look at it. If something could make data research sexy, this is it.

Google Killer?

Now, even though Alpha does inherit visual cues from Google — the single-line entry field and the ability to accept regular queries in English for example — there’s little in common between Google and Alpha. Initially touted as latest in the long line of Google killers , Alpha is very different from Google — it computes the answers to questions that may not even have ever been asked, based on mountains of organised data.

In contrast, Google primarily directs you to Web pages where other people have already created answers to questions. It’s a very different type of service, though there may be areas of overlap.

The trouble with Alpha is, if you venture beyond the data sets that Alpha currently contains, it just gives you an unhelpful error message: “Wolfram Alpha isn’t sure what to do with your input.” At this point, it doesn’t even help you find your way back to queries that do work when you’re lost as you will often be when you’re looking for data but unsure of what words to use to query for the data.

Due to its deep logic and structure, it will make demands on users — and in order to use it well, people will need to learn some new concepts and a query language, which may well be too much to ask of your grandparent who’s just about gotten used to Googling. It’s just not mainstream enough.

Incidentally, Alpha will come in a free version, but there will also be a paid version, which will allow users to download and upload data to Alpha. Stephen Wolfram hasn’t gone into too much detail yet about pricing, but pro users will, for example, be able to not just see a graph, but also download the data behind this graph for use on their own machines or in the Mathematica software.

And, as it opens to the public, it’s more and more likely that initial users will find the hard edges of its capabilities. Try your past 10 Google queries on the day Alpha launches, and it should be interesting to see how it compares and whether it trips up. The real strength and power of Wolfram Alpha does seem to be for the academic and research community, and as an entertaining exploration tool for those with an interest in the sciences.

Its corpus of information is undoubtedly of much higher quality than the Internet information available to Google. However, despite having such a lot of data, this is still a tiny fraction of the information on the Internet, so it is unlikely to have information on some random topic that you are interested in.

Even with the academic community, its success will ultimately depend on its ability to interpret user inputs. Public demos with canned data are one thing, but the queries of real users are another.

If anything, Alpha will probably be a worthy challenger to Wikipedia and many textbooks and reference works. Instead of looking up basic encyclopaedic information there, users can just go to Alpha instead, where they will get a direct answer to their question, as well as a nicely presented set of graphs and other info. Google Killer? Not in the least. Complement to search engines — most certainly. Potential Google acquisition — quite possibly!

Other Google killers

To be fair to Alpha, it did outdo Google in one little-known aspect. On the same day when Wolfram demoed Wolfram Alpha, Google launched a new system that visualises public data. If you search for a US state’s unemployment rate or population, Google shows an answer much like Alpha’s, replete with graphs change rates over time. Now, search is an altogether different ballgame, and Alpha’s not toppling Google off the top of the hill just yet. History bears witness to a number of has-beens who’ve been trumpeted as next Google on launch, but have faded into history thereafter.

Like Mahalo, “the world’s first human-powered search engine”, Quintura, the “visual search engine”, or Spock, “the people search engine”. All three offered niche search capabilities that in no way matched Google’s coverage and ease of use. Spock suffered at the hands of FaceBook too.

And then Cuil (pronounced kewl) came along. Founded by ex-Googlers, Cuil claimed to index far more of the web than Google, but had disappointing and irregular results. Reminds me of the ’90s, when Barnes & Noble tried to catch up with Amazon by offering more books, not a better experience. Didn’t work then, doesn’t work now. The biggest challenge for Cuil and the other would-be Google killers is rather simple — they first have to offer a search engine that works as well as Google’s, then beat Google. Not happened so far.

But then again, the search market is just too big, and too lucrative, for the Cuils of the world to pack it in this early.

Top
Email This Page