Just as British MPs line their pockets with taxpayer cash, just as Hollywood epics re-write history with happy endings, so Google was always going to want to compete with the Wolfram Alpha computational engine head on. And so it is that we get a sneak look at Google Squared in the Google Labs, and answers engine rather than a straight search one.
So what is it, how does it work and perhaps most importantly is it any good?
What it is, in the words of Google itself, would be an "experimental search tool that collects facts from the web and presents them in an organized collection, similar to a spreadsheet." More similar to a grid, or funnily enough given the name, a square. Instead of returning web pages it returns facts. Searching for US presidents provides grid containing at a glance information including dates of birth, full names and additional data. Each square on the grid is interactive, clicking opens up the web page which was the source of the information contained. Ah yes, that brings us nicely to the how: whereas Wolfram Alpha provides answers from its own database of information, Google Squared gets its data from searching the web and scouring the pages there for the relevant stuff which then has to be analysed, sorted, filtered, condensed and displayed within the grid. If you want additional information, you can add another square (or column if you want to stick to the spreadsheet analogy) and it will endeavour to go get more for you and fill that square in.
The problem being, especially if you thought that Wolfram Alpha made you work hard trying to get the answers that you wanted, is that it is a rather hit and miss affair. Don't bother ego surfing, it won't know who you are. Actually, it won't know much about anything unless you choose one of the 'suggested searches' that I suspect have been well researched up front. Google has an answer of course, that the "technology is by no means perfect" and that is why it has been designed to be "conversational."
What? A conversational search engine? If I wanted conversation I'd ask my wife what she knew about the sexiest actresses on the planet or which beer is the best. Google reckons that Squared enables the user to "respond to the initial result and get a better answer." Great, here's a thought Google: how about you give me an answers engine that provides the answer to the question I ask instead of acting dumb and asking what I really meant.
Aha, Google says, but while "gathering facts from across the Internet is relatively easy (albeit tedious) for humans to do, it's far more difficult for computers to do automatically."
It would also appear, from both Wolfram Alpha and Google Squared, that building an answers engine that actually works is really rather difficult. By which I mean building an answers engine that works in a way that the general public, you know the people using it, can understand and returns answers to the questions they ask rather than questions the search provider wants them to ask.