- Saturday, 21 November 2009
- INQ Mobile
- RSS
There was an immeasurable distance between the quick and the dead: they did not seem to belong to the same species; and it was strange to think that but a little while before they had spoken and moved and eaten and laughed - W. Somerset Maugham
Google's new search offerings seek to impress
IF GOOGLE WAS SEARCHING for a way to impress, one would have to look no further than last night's 'Searchology' event, hosted by the firm in it's MountainView headquarters to show off its coolest innovations and enhancements in search.
Declaring that "search is still in its infancy," Google's vice president of search products and user experience, Marissa Mayer said the company could not rest complacently as "the race in search is far from over and innovation and continued improvement is absolutely pivotal".
Mayer said Google's engineers were "worried" about what the next big thing in search might be and how they'd find it before anyone else did. Naturally, the buzz word "semantic web" was also tossed around for effect.
Meyer said her firm had been picking up its development pace, having released 365 products in total last year and already 120 in the first quarter of 2009. To illustrate her point, Meyer showed off several cool new offerings to wow the crowd.
First up was something Google is calling "Rich Snippets", search results which offer users a quick way of getting to the information they're looking for. As an example, if a user was googling a particular restaurant, s/he would be able to get a "rich snippet" showing the establishment's review scores, price range, address information, and so on, all culled directly from the restaurant site's metadata.
Next, Meyer showed off Google Squared, an offering set to go public in around one month's time. The app, which Meyer gushed was "transformative," can yank information from across the web and whack it up into a spreadsheet in "split seconds."
Giving a doggy style example, Meyer showed how a query for "small dog" pulled up a table complete with photos, origins, weight and height of various breeds in seconds.Meyer said this had been achieved by analyzing structures on the web which seemed to imply facts, then "corroborating the evidence around whether or not something is a fact by looking at whether that fact occurs across pages."
Coming up with such a table, however, was no easy task, admitted Meyer, noting "it takes an incredible amount of compute power to create those squares".
Next up, Meyer showed off Google Search Options, which lets Googlers "slice and dice" search results in such a way they can get results by genre, that is, product reviews, forum posts, videos, blogs, images, timelines etc.
Yet another cool search feature is Wonder Wheel, which spins up a visual concept map highlighting related topics. In an example, a search for the name of the Los Angeles pro basketball team the "Lakers" rolled out a visual wheel with spokes including "Lakers roster," "Lakers rumors" and "Lakers Rockets."
Last, but not least, Meyer got the audience all starry eyed over an Android mobile app dubbed Skymap, which uses a phone's GPS, compass and accelerometre to beam up a dynamic star map which knows your exact position. The app is apparently already available from today on the Android app market for all you nerds searching for something to do with your free time at night. µ
L'Inq
BBC
Share this:
ShareBusiness Intelligence Software
Customer Relationship Management (CRM)
Enterprise Accounting Software
Enterprise Resource Planning (ERP)
Enterprise Systems Management
google is having the touch. the rest are backpackers with heavy load.
Many commentators give the impression that Google has some sort of magic AI that can read a document, determine it's context and place it in a pile, with other, similar, examples. In fact, they achieve a lot of this by analysing meta-data explicitly embedded around the content, by web developers, who deliberately write the content, or the XML-feeds that drive it, in this way, for exactly these sorts of reasons. In this case, we know, it's microformat data, that they're using; but we don't know precisely how they're using it. We know that semantic web design, where the document embeds data that describes its content - we know this is a Good Thing, for many reasons: Google pagerank ebing one of them, but we don't know exactly how any of it helps.
Google exists in a love-hate relationship with this process, wanting it to develope without explicitly saying, for instance, "our bot looks for divs with ids that contan the word 'address', when looking for address data, on a page" or "our algorithms seek out definition lists and instances of the abbreviation tag". They don't explain the exact process, since it would subvert their advertising revenue: page ranking would become a matter of good design, rather than dollars. It would also make it easier for bad guys to game the system, by pushing their phishing websites up the page ranking to lure the unwary - as many of them already do, by querying Google's own 'trends' system, to automatically find out what topics are 'hot', and gearing their phishing pages accordingly.
And so we end up in a world where The Search Engine appears to magically work things out for itself, while web developers frantically try to reverse engineer the process, and determine what it is, they're doing right, to keep the Googlemonster happy ("Hey, we're running kinda low on virgins... has anyone tried feeding the dragon some fat schoolboys, recently?").
Strangest of all, though, is that fact that, if you look at the Google cache for this page, you'll see that the Googlebot had already read, ranked, rated, and stacked, the contents this article, twenty three minutes after Sylvie pressed 'Publish'.
looks like sylvie is doing a new thing at l'inq: double meaning-ed expressions and phrases; doggy style ... et al.