This article is more than 1 year old

Google crossbreeds search with spreadsheet

Semantic squares

Searchology Google is breeding a newfangled search tool that automatically organizes web data into the familiar rows and columns of an ordinary spreadsheet.

That may sound trivial. But in Googleland, it’s close to walking on water.

Dubbed Google Squared, this experimental tool will make its public debut as a Google Labs offering "later this month." Tuesday morning, at the annual Searchology dog-and-pony show inside its Mountain View headquarters, Google demoed the beta for gathered journos.

In typical fashion, Google VP Marisa Mayer was nearly breathless in describing the company’s latest search prototype. "For a long time here at Google, we’ve known that users often do deep research on the Google search engine, keeping track of different entities and values of those entities, ultimately building a spreadsheet," she said.

"One of the hardest problems in computer science is data abstraction - looking at the unstructured web and abstracting values and facts and information in a meaningful way in order to present it to users, building out some of these research spreadsheets in an automated way. But that’s no longer a hypothetical."

With Google Squared, project manager Alex Komoroske explained, you could conjure a spreadsheet detailing various breeds of small dogs. It could automatically visit the web for, say, descriptions of each breed, its weights, heights, photos, and so on - before tossing the lot into neat rows and columns.

If you then mouse over individual squares of information, a pop-up shows where the info came from.

How does it work? Well, it doesn't always work. It might suit, say, astronauts or US presidents. But an auto-spreadsheet detailing, say, the films of Ernest Lubitsch would likely be a step too far.

"If it's not a topic that a lot of people have talked about, it's going to be somewhat harder for us," Komoroske told The Reg. "With an example where lots of people have discussed a topic - created information about a certain topic - we'll do pretty well."

Clearly, this involves some sort of contextual analysis of web pages - semantic search, if you will. Asked about the details, Mayer was happy to provide an inside view of the project, which was incubated in Google's New York office.

"I think we can open the kimono a little bit," Mayer volunteered. "I will say the technology behind it is totally amazing."

And naturally, she said that without a hint of irony. This is Google. ®

More about

TIP US OFF

Send us news


Other stories you might like