Fight! Google Squared vs. WolframAlpha

By now we all realise that WolframAlpha is not intended to compete with Google's Universal Search; it's a 'computational knowledge engine' designed to serve up facts, data and scientific knowledge and is an entirely different beast. Nevertheless, Google is not a company to be outdone and has just announced the release of Google Squared which, if the technology press is to be believed, is Google's attempt to usurp WolframAlpha's grip on offering up facts, data and knowledge. Indeed, Google attempted to steal WolframAlpha's thunder by announcing that Google Squared was in development on the same day Stephen Wolfram was unveiling WolframAlpha for the first time a few weeks ago. Meow!

In the same way that WolframAlpha occupies a different intellectual space to most web search engines, Google Squared seems to be quite different to WolframAlpha. Says the Official Google Blog:

"Google Squared is an experimental search tool that collects facts from the web and presents them in an organized collection, similar to a spreadsheet. If you search for [roller coasters], Google Squared builds a square with rows for each of several specific roller coasters and columns for corresponding facts, such as image, height and maximum speed."

Google Squared appears to work best when the query submitted is conducive to comparing species of, say, snakes or country rock bands. With the former you retrieve a variety of snake types, images, description, as well as biological taxonomic classification data, and with the latter genre and date of band formation is retrieved (including Dillard & Clark and the Flying Burrito Brothers), in addition to images and descriptions. Many of the data values are incorrect, but Google has been quite forthright in stating that Google Squared to extremely experimental ("This technology is by no means perfect"; "Google Squared is an experimental search tool"). Of course, Google wants us to explore their canned searches, such as Rollercoasters or African countries, to best appreciate what is possible.

As we noted recently though, place names are good to test these systems and, like WolframAlpha, some bizarre results are retrieved. A search for Liverpool seems to only retrieve facts on assorted Liverpool F.C. players, and Glasgow retrieves persons associated with Glasgow and the death of Glasgow Central train station in 1989(!) I had hoped Google Squared's comparative power might have pulled together facts and statistics on Glasgow (UK) with the ten or so places named Glasgow in the USA and Canada. A similar result would have been expected for Liverpool or Manchester (which has far, far more), but alas. This is a particular shame owing to the fact that much of this data is available on Wikipedia in a relatively structured format, with disambiguation pages to help.

Google Squared allows users to correct data, remove erroneous results or suggest better results. The effect of this is a dynamically evolving result set. A search for a popular topic an hour ago can yield an entirely different result an hour later. All of this will help Google Squared become more accurate and cleverer over time.

Although Google Squared and WolframAlpha are quite different, there are some similarities. For this reason it is possible to state that the current score is 1-0 to WolframAlpha.