Speeding Up the Web
The web is transforming from its earlier days when every bit of information was carefully considered by a webmaster before adding an image or some text to a web page.
I remember spending hours and hours optimizing images so that they were small and still decent looking, and squeezing white space out of html to make pages faster for phone modem transmissions. Learning as much as possible about Cascading Style Sheets was important because they could help shave a lot of html out of a page. Not everyone considers this stuff, and it feels kind of odd going to sites that get millions of views a month, and seeing them using tables and font tags.
Bandwidth has improved, and more images, pictures, video, and content comes across the screen than ever before. And, that kind of use will probably only grow. Which means that we will need more bandwidth.
Google’s Need for Speed
It looks like the search engineers at Google have been thinking a lot about reducing the amount of bandwidth and processing time for queries as much as possible, too. What will this mean for searchers, and for web site owners? That’s a question worth exploring, and worth considering carefully.
A new patent application from Google looks at ways to anticipate searches from users, and start responding to their queries before those are even fully made. And to start answering followup queries before they are even needed.
Why is this important?
For one thing, bandwidth through handhelds and wireless doesn’t deliver the speed that a faster connection might, and those methods of delivery of web content are increasing tremendously rather than diminishing. Wireless transmissions and reception also can be adversely affected by different forms of interference, so quick returns on searches are essential. Here’s the patent application:
Predictive information retrieval
Invented by Shumeet Baluja and Henry Rowley
US Patent Application 20060122976
Published June 8, 2006
Filed December 3, 2004
Abstract
A computer-implemented method for generating results for a client-requested query involves receiving a query produced by a client communication device, generating a result for the query in response to reception of the query, determining one or more predictive follow-up requests before receiving an actual follow-up request from the client device, and initiating retrieval of information associated with the one or more predictive follow-up requests, and transmitting at least part of the result to the client device, and then transmitting to the client device at least part of the information associated with the one or more predictive follow-up requests.
What kind of uses might predictive queries cover?
We have looked at a couple of other uses of predictive queries here before, including one use involving hand held devices, and another one used in applications like Google Suggest.
But, what about other searches from Google?
What about web searching, shopping searching, news searching, image searching, or a combination of those? Would it help the search engine to anticipate queries for purposes of deciding what advertising to show next? Those are the types of searches that this predictive queries patent application may cover.
How might predictive queries work?
- A person makes a search request,
- A response is generated in response,
- A request predicting module attempts to determine one or more predictive queries as follow-up requests by that user, and;
- That follow up also triggers one or more predictive follow-up requests before they are made, and has them ready to display.
The system might base predictive queries upon such things as recent requests from others for the same or similar information, or statistics aggregated over a number of users:
For example, people who query “coffee” may frequently also query “Starbucks,” so when a particular user enters “coffee” as a search, the system may pre-fetch results for the search term “Starbucks” while the device is displaying to the user the results of a search on “coffee.” Then, if the user acts consistently with his or her peer group, the follow-up information is nearly immediately accessible.
It sounds like the Wisdom of Crowds may be what Google relies upon to anticipate what people will search for next.
4 thoughts on “Google predictive queries”
Comments are closed.