Xerox's PARC Gets Serious
June / July 2008
Volume 6 Number 3
The Xerox Palo Alto Research Center is famous for fumbling the future. In one of the great mistakes in corporate history, PARC's researchers developed many of the fundamental technologies behind the personal computer but failed to commercialize them. Steve Jobs visited the vaunted research campus in the hills above Silicon Valley and walked out with some of the best ideas, such as the computer mouse. He used the inspiration at Apple to jumpstart the personal computer revolution.
Xerox had an excellent copier business at the time with high margins. It just didn't see the personal computer as vital to its core business. Mark Bernstein, the current director of PARC, is quite familiar with those past decisions.
"It generally comes up in every interview we do with journalists," Bernstein said. But Bernstein is confident that PARC is now structured to both innovate and commercialize its R&D. He argues that the research center has found a balanced business model for surviving in an age of corporate R&D cutbacks.
"We're working on important problems that are commercially valuable," Bernstein said. "The end game is about commercialization."
Xerox spun out PARC in 2002 as a separate company. At the time, it was funding almost 100 percent of work under way at PARC. Now, about 50 percent of the funding comes from the outside from entities such as a joint venture between Xerox and Fuji. About $10 million a year comes from government funding sources.
Bernstein says that licensing and other fees more than cover the costs of PARC's 165 researchers. And projects live or die based on how long it will take to get products out the door.
To promote commercialization, PARC has eight business development staffers embedded in its research group. Those people are close to the research and their job is to get others outside the company interested in the projects. PARC also has a program called Startup@PARC where it invites startups to incubate within PARC. The record over time, in spite of the PC mistake, has been good. More than 40 startups have spun out of PARC over several decades. And Xerox made a lot of money off the laser printer, another of its legendary inventions.
Even as a spinout, PARC is still a very important part of Xerox's emphasis on R&D. Xerox has more than 55,000 patents and still gets 10 a day, according to Sophie Vandebroek, Xerox's chief technology officer. The company invests $1.5 billion a year in R&D and has more than 800 corporate R&D researchers worldwide. Their work is focused on documents, mass customization and environmental technology.
With its more focused approach to commercialization, Xerox is still criticized for failing to pay enough heed to basic research. As in other quarters, researchers who are doing fundamental work still have a hard time getting funding.
"We've been able to get to a balance that is more grounded in the reality of today," Bernstein said. "We have always had an ethnographic view of technology, with a focus on things that you can use."
That balance between fundamental research and commercialization shows in Xerox's efforts over several decades in linguistics research. That has more recently evolved into research into the semantic web. Bernstein says that the next revolution in information will come from semantic search, where the search engine understands the meaning of what you want and the context in which you're searching for something. Such searches can deliver much more useful, filtered search results.
An example of this is FactSpotter, a semantic technology developed by Xerox researcher Christopher Dance. It goes farther than a typical search engine because it analyzes the semantic context of text to infer whether ambiguous words are being used as nouns or verbs. It also finds the person or object to whom a pronoun refers. If an article is about Steve Jobs and it mentions "he" a lot, the FactSpotter engine will note that the article is indeed full of references to the Apple CEO. Search engines that don't use this technology would have less useful results.
Dance has also focused on a hybrid of text and image searching in order to deliver smarter results than separate text and image searches. The hybrid of visual and text databases results in a search engine that can take your diary from a trip to Peru and combine it with the pictures you took. When it comes time to building a personalized blog, you can use the technology from Xerox to automate the process. For instance, if the word Cuzco appears in the text, the system will review your photos of Peruvian cities and then conclude which photo fits with the words.
Xerox's labs are full of such projects with potential to turn the world as we know it upside down. Another is "eraseable paper." Researchers Eric Shrader of PARC and Paul Smith of Xerox's Toronto lab showed how they could save a lot of wasted paper printouts by printing on the same pieces of paper over and over again. These researchers created their eraseable yellow paper with photochromic switch materials. This is the kind of light-sensitive material that is used in sunglasses that darken in the sunlight. That means they change color when they are struck with light, in this case ultraviolet light. The UV light bar moves up and down a page and shines its light where needed to make an image on the paper. A printed page fades within 24 hours.
Other document and image technologies are in the mix, like "higher dimensions of documents," a project aimed at letting you view what a document will look like in a 3-D view before you print it. Another is solid ink printing, where you melt a solid block of ink in order to print an image on a paper at a high speed.
But not everything has to do with Xerox's core business. A Fuji Xerox research project is focused on how to replace plastic parts in copiers with biomass materials. It has begun testing such a corn-based plastic replacement material now, though the cost is three times as much as plastic.
Along those same lines, Richard Bruce is operating on the fringe of Xerox's main corporate initiatives. He is using modern DNA analysis techniques to find cancer cells at an early stage and to do important genetic tests on fetuses in a way that doesn't endanger a pregnancy the way that amniocentesis does. The techniques for doing these tests today are either inaccurate or slow. Bruce's technique borrows from laser printing. But if it hits paydirt and is used to prevent the 1,500 amnio-related miscarriages that happen every year, Xerox could license it out.
Some people might view Bruce's project as research gone astray. He might be better off working for a biotech company. But this is a case where Xerox sees the potential commercial value and allows the research to go forward, even if it is not core to the business. It is such cases that may allow Xerox to capitalize on breakthroughs and generate value from research outside its core markets, Bernstein says. And that could be a way for Xerox to extract value from the next PC-like revolution, rather than fumble it.
Dean Takahashi is lead writer for digital media at VentureBeat. He can be reached at deantaka@gmail.com.
Copyright © 2012 | Innovation America