Programming the search engine
As we reported awhile back
, Google should introduce a new technology later this year, called the Programmable Search Engine, that could significantly change how Internet search is done. Reading
over some of
the patents around the PSE
Google was recently issued offers a glimpse into this new world.
What Googzilla wants to do, in a nutshell, is bring a bit more precision to search when it comes to vertical sites. 'Vertical site' is a marketing term for any specialty Web site with a lot of material, such as Amazon.com
, or the Environmental Protection Agency
or, for that matter, Government Computer News.
These days, such sites can offer a wealth of information on a particular set of topics, though in most cases, the search engines these sites offer are still pretty sketchy (ours included). You can find what you are searching for'and a little more.
For instance, if you go to IRS.gov to search
for a Form 1040x you may need, you will get back a link to a PDF version of the form, as well as a number of other links which may or may not be useful to you.
Google wants to tighten that up. Imagine, in the future, if you look for the 1040x, you get back a list of options that clarify what you are truly seeking. Do you want to download the form? Do you want instructions on how to fill out the form? And so on. In fact the site may already have a pretty good idea of what you need, based on your history of surfing and preferences submitted by the site itself.
'When the user issues a query associated with a particular context,
either by clicking on a particular link or performing the query from a
particular vertical content site, he or she is effectively communicating
a piece of information about his or her current activity and/or about
him- or herself,' the PSE patent reads.
What happens in the background for these types of searches is that the organization's Web site hands the query off to Google, along with a file that may offer some insight about what users really want when they submit searches for terms like 'Form 1040x.' Google may have its own files that could help further clarify the context of the query as well.
'The pre-processing operations include operations to revise, modify or expand the query, to select one or more document collections on which to
conduct the search, to set various search algorithm parameters for
evaluating the query, or any other type of operation that can refine,
improve, or otherwise enhance the quality of the user's search query," the patent again.
Of course, this service will require some work
on the part of the vertical site manager to build these files. And as agency Web site traffic increasingly comes from Web sites (Google, mostly)
, managers should consider the full implications of having your information so finely parsed and presented to the outside world, as other 'vertical' fields of endeavor are starting to do.
Posted by Joab Jackson on May 30, 2007 at 9:39 AM