Search Engines of the Future

Search results may also be narrowed through the use of tagging and file-sharing software. In these cases, visitors to a site mark it or comment on it. This community of like-minded people thereby produces a set of results narrowed by their common interests. This process of tagging Web sites is called folksonomy—“folks” plus “taxonomy.”

Local searches, or Web searches that are confined to results that have a physical location in common, are predicted to become much more important in the coming years. Local searches involve the ability to inform your computer, cell phone, or PDA that you're in ZIP code 11743. Suppose you want a pizza but you don't know where to place an order? Instantly, you have all the data that you need to order a pie from any of the restaurants within your ZIP code. This is present, not future, technology and, increasingly, at least some of the computers already know where they are through personalized settings. A cell phone or PDA linked to the Global Positioning System (GPS) knows where you are by satellite.

One of the limitations of a conventional search is that it is heavily dependent on HTML—the hypertext markup language that makes links work. But HTML is based on words, and that makes search engines clumsy when it comes to retrieving pictures or videos or music. Of course, if they are tagged, the crawlers can index them. Flickr, a Web site that indexes hundreds of thousands of photos, has access to sort those images, but the tags are verbal.

Recent research into search engine technology involves visual searches. Scirus, a search engine for scientific information, allows users to draw an image—even a 3-D image—with a mouse and then search for similar results in its database but not yet within the Web. After all, engineers and designers may be searching for a shape, not a verbal description, and “sort of looks like a hammer, but pointy” is not going to return images of a hatchet. Search engines that utilize visual imagery remain limited, but the technology is already here and it may be available on a wider basis in the future. Search engine creators imagine that users will one day be able to search for an exact visual image that they remember from a movie or a photograph. Other similar technology will allow users to play a few bars of a song or a piece of music on a virtual keyboard in order to access a complete recording of the entire song.

One of the fantasies in the world of search is that of mobile visual search technology. John Battelle imagines pointing your cell phone at the bar code of a product in a store and receiving information about it—everything from user reviews to whether the product was made using child labor—including where in the neighborhood you might buy it less expensively. This fantasy requires your cell phone to read bar codes. It also requires all merchants to make the databases of their inventories searchable. It remains future technology, but technology that will likely be achievable within the next decade.

But another Battelle prediction that is coming true is the use of Mobot, a visual search technology that utilizes a camera phone and any wireless carrier. Instead of typing on a keypad, the user takes a picture of an ad, a product catalog, or a magazine and immediately connects to information about it, without URLs or bar codes. Battelle imagines, however, that we will one day have the technology to point a camera phone at anything and receive information about it—in a museum for instance.