Reporter's Notebook: A top Google engineer gives a rare sneak peek into how the search site perceives and uses PDF documents.ORLANDO, Fla.—To outsiders, Google as a company seems to have retained the quirky flavor of a small Internet startup despite running the world's most visited Web site.
But creativity still happens as in the old days: Marissa Mayer, Google vice president of search products and user experience, recounted in her May 10 keynote address here at the Aquent Graphics Institute's Adobe Acrobat & PDF Conference how she and Google founder Larry Page themselves hacked together the system of converting the first paper book into a Google Book, downloadable as PDF files.
First, Mayer and Page set up a hi-res camera, held in place with rubber bands. She flipped the pages while he ran the shutter. Yet it wasn't quite working out until they broke out another piece of hardware not normally associated with photography: a metronome, the device musicians use to keep steady time while playing their instruments.
"We [took the pictures] to the beat of the metronome, so he wouldn't be taking pictures of my thumbs," said Mayer, an artificial intelligence specialist who was one of Google's first employees and did much of the early development of the search engine's interface.
Since that fateful day, Page, Mayer and a legion of Google believers have uploaded more than a million volumes to Google Books. Mayer, who calls the project "Google's moon shot" at making the most well-researched information available online, says the site continues to work with libraries and publishers to expand the titles it offers.
Click here to read more about Foxit Reader, an alternative PDF application.
Half the company's new innovations—and other services the site offers, such as Google News and features of Gmail—come during what the company calls "20 percent time," or the fifth of one's office bandwidth the Google sets aside for employees to work on pet projects that interest them.
Google is fairly format-agnostic when it comes to documents, preferring to concern itself more with cataloging what Internet users are interested in. So that means when PDF joins MS-DOS in the great techno-graveyard in the sky, Google will be searching for its replacement and, probably, rendering it in HTML or offering links to direct download as it does with PDFs today.
Yet, agnostic or not, Google has to acknowledge that PDF dominates the Web when it comes to documents meant for offline storage, viewing and possible post-download printing. So how does Google perceive PDF as a technology and build that technology into its own operations?
Mayer said in an interview with PDFzone after her address that Page lobbied hard for several months to make Foxit Reader an element of Google Pack, the company's basic utilities download that enables PC users to quickly load new machines with the software they need to use Google services. Eventually, Google signed a deal with Adobe to make Reader the PDF viewer in Google Pack, despite Page's concerns about its load time before it went live in 2006.
"PDF is essentially a platform at this point," Mayer said. "One thing that has pained both the search side and on the [Google] Pack side is the startup time of the app. That's what causes us to do 'view as HTML.' The size of the application and the speed."
She added that although Gmail now can render PDF attachments in HTML, users shouldn't look for Google to do the opposite—outputting e-mails or other content as PDFs—soon. Given the scale on which Google works, she said, that would represent a massive undertaking that the site is not prepared to take on at this point.
Don Fluckinger is a freelance writer based in Nashua, N.H., who has covered Acrobat and PDF technologies for PDFzone since 2000.
Document Management 
