Google is not the obvious company to telephone when you are looking for directions to a restaurant or hotel, but the popular search engine's development team is hoping that its emerging voice search facility may over time completely change the concept of a search engine.
Craig Silverstein, director of technology at Google and the company's first official employee, was in London on Tuesday and spoke to ZDNet UK. He was at university with Google's founders, Larry Page and Sergey Brin, and helped transform the idea from a research project at Stanford into one of the hottest Internet properties. Now he's focused on the future of search and is excited about some innovations which are accessible in pre-beta form in the Labs section of the Google site. He said that although the utilities are not ready for prime time, they gave an indication of where the company is heading: "Sometimes we risk our reputation if these projects are out there and have some problems, but we think the user's benefit is more important than some nasty coverage," he said.
Searching in the supermarket
Silverstein said he believes that within a few years Google could have a voice interface for everything from driving directions to help you finding the aisle for a particular food in your local supermarket.
"That's something you would never think to ask a search engine. You're not likely to be using your laptop in a supermarket, but in the future I think search will be far more accessible -- you won't be tied to your desktop, you will be able to do it from your mobile phone or PDA -- and you'll start to see search used in fundamentally different ways. The kinds of things people want information about when they are walking around or sitting in a bar is very different to what they want while they're at home," he said.
An embryonic version of the speech search is already available on the Google Labs Web site, where the company lets Web users play with its pre-beta projects before they are ready for "prime time". Silverstein said speech-based searching presented a real problem but not because of the recognition technology. Instead he said the problem lies in the way results are returned.
"The problem is, how do you get the answers back? Do you have someone reading them off to you like one of those voicemail mazes where it takes so long to speak to someone? A big list works visually, but doesn't work very well in audio," he said. For now, Google's Voice Search can receive keywords by phone, but it displays the results in a browser window.
Silverstein said the problems presented with audio searching will be solved as computers improve: "When computers can pick the information that is really useful, as opposed to displaying a big list and letting you decide what is useful and what isn't. Already we've seen a sea change because information is much easier to find than it was before the days of computers. I think we'll see another one after which information will be easy to find wherever you are," he said.
Bomb the Google
Silverstein said he was unconcerned about the phenomenon known as Google Bombing. Human nature dictates that Web searchers are more likely to click on the results at the top of the results page rather than scrolling down the list, so Web administrators and marketers have been keen to try different ways to improve their chances of being discovered in a search. One such method is called Google Bombing, where Web sites add a carefully worded hyperlink pointing to a specific Web site in an attempt to boost that Web sites ranking when the phrase is part of a search.
A recent example of Google Bombing was when anti-war protesters successfully made the first result for "miserable failure" point to George Bush's official biography. Pro-Bush sites retaliated by Google Bombing the same term, but pointing to one of Bush's high profile critics, Michael Moore. Within weeks, Moore's official Web site was ranked second on the list. And the battle continues.
Silverstein admits it is possible to affect Google's search results but he said he didn't consider it a serious problem, especially when there isn't an obvious search result in the first place: "The term 'miserable failure' doesn't have an obvious result and it is going to be fairly random to begin with. In cases like this, even if a fairly small percentage of the Internet community do it, the result will pop up. But for something where there is an obvious good result -- like the IBM home page -- it is not going to be possible," he said.
Searching beyond text
Silverstein said that Google is also interested in adding new kinds of content that were not previously available in any electronic form. "I think we probe much if not most of the static Web, which consists of pages that are not dynamically generated. From the dynamically generated stuff it is hard to say. We cover a lot of it, but probably not close to everything," he said.
One area where the company has added new content is in its Google Catalogues section, in which entire shopping catalogues have been scanned and published: "We took a bunch of mail order catalogues, many of which are not online because they are very small. We converted them to text and made them searchable. This information wasn't even available electronically but now you can search it and we are hoping to get more of that type of information available," he said.
Talkback
delivering new content, what about smell. Blueyounder, Tiscali (UK) company have developed a pice of hardware that mixes scents so that when you open an email it could mix a chocolate smell from a company like Torntons.
I would like to search my supermarket, as predicted in the article and smell the product, view it in 3D, get associated products - in fact have recipie(s) attatched.
Though it would be a dream come true, that a single source can answer every single question, we may have something like an Oracle, except more powerful. But I doubt google will have the rights to do that.
Google is so popular that sites are sustained by the traffic sent to them via this gigantic search engine. Once google takes away that, ie. giving users no reason to go to their site, then these sites will have to shutdown, or insert in their robots.txt that google cannot spider their pages anymore. Sites need the traffic so they can sustain themselves (via advertising or selling products), but if someone creates a wall that is appealing and is a buffer not to go to another site, these smaller sites which rely on google and other search engines to drive traffic to them will crumble, which will result in google crumbling very slowly.
Google would not have existed today if it weren't for other sites allowing them to be spidered. They cannot take advantage of that.