For Xerox Corp. (XRX ), the millennium got off to a rocky start. Revenues have fallen every year since 2000, and Xerox doesn't expect to reverse that trend until the latter half of 2004. But the Stamford (Conn.) company is confident that new technology will restore growth and profits. Certainly, its research operations have a legendary track record for innovation. The Palo Alto Research Center (PARC) hatched many silicon-age icons. In 1973, almost a decade before Apple Computer Inc.'s Macintosh, PARC built the first graphical-screen personal computers -- and soon after linked them via an Ethernet network. Xerox famously failed to profit from most of these early inventions, except for the laser printer. But lessons were learned. Recently, for example, Xerox has scored big with a laser-printer advance that applies four colors simultaneously instead of sequentially.
The man recharging Xerox' research and development is Hervé J. Gallaire, chief technology officer and senior vice-president for research and technology. At 59, Gallaire has been Xerox' CTO for four years. He oversees 1,000 scientists and engineers at labs in Webster, N.Y., and in Britain, Canada, and France -- plus PARC Inc., now a separate subsidiary of Xerox. Before joining Xerox in 1992, he had been head of computer science and mathematics at France's National Superior School for Aeronautics & Space and managed hardware and software development at French computer giant Groupe Bull. To glimpse what's in the new-technology pipeline, Senior Writer Otis Port chatted with Gallaire in New York.
Has the turmoil affected your job much, especially the spin-off of PARC?
Not really. My job is still doing the work that needs to be done to prepare Xerox for its future, both in terms of improving what we do today and in terms of finding new types of products and solutions. We set up PARC as an independent company, but this doesn't mean Xerox is any less interested in our strategic relationship with PARC. Some of the technology that comes out of PARC is absolutely essential.
But we also recognize that PARC can do much more than just serve Xerox' needs. The technologies they develop could benefit others as well -- and should. So the intent is to provide a set of partners that sponsor work at PARC, and eventually some of these partners could become investors and partial owners. PARC already has some strategic research partners but not yet investors.
What would happen if a potential competitor wanted to invest in PARC?
We would have to look at the specifics. We probably would not do a deal with, say, Canon (CAJ ). But if we wanted to use a new PARC technology in Xerox printing or computing products, it would be O.K. for somebody else to use it in a market like bioengineering.
What are some of your hot technologies?
Historically, Xerox has had a big impact on how people work. A lot of the business processes in offices today have been really affected by printing and copying. But there's still more to do.
A Canadian study doneestimates office workers spend about 55% of their time dealing with documents in one form or another -- electronic or paper. But people are not very productive at this, since they say 40% to 66% of that 55% is wasted time.
We have social scientists and anthropologists who develop processes for extracting information about how people work. You can't just ask: "How do you do this?" You need a methodology. We spend research money on these kinds of tools -- and related ones that help people and offices improve the way they work. With our optimization tools, they typically reduce costs by 20% to 30%.
How does that affect office paperwork?
The paperless office is a myth that's probably never going to happen, so we want to help people become more efficient at using paper. We are developing a technology, centered around scanning, that extracts information from documents so we can index them better, archive them better, and retrieve them better. In fact, we're using one of these technologies in-house today. It's called CopyFinder. If you put a page of paper on a copier, the copier will associate it with similar documents in your personal repository, in your work group's, or even on the Web.
How does it do that?
By scanning the document, performing OCR [optical character recognition] to digitize the text, and then doing content analysis. Using automated linguistic tools developed in Grenoble, France, and at PARC, it determines the key characteristics of that page. That content analysis becomes the basis of a search for matching documents. This is still in research.
If the copier is connected to a network, can it send e-mail?
That's called "scan to e-mail," and a lot of Xerox people and customers are doing it. They also scan-to-fax and scan-to-archive. It saves a lot of time, even without the linguistic tools for classification and searching that we have in the lab. Law firms are realizing huge benefits from this, because so much of what they get is in paper form.
For some big customers, we provide scan-to-archive as a service from a site in Arkansas. One financial firm archives 5 million documents a year, and we maintain their archive. Now, from their offices here in New York, or anywhere, they can access a document in a half-day or less. Before, it took a week or two to get a document out of the paper archive.
How much do you spend on basic research?
Overall, Xerox spends close to $1 billion on both R&D and R&T [research and technology]. Basic research is about 20% of my budget -- not 20% of overall Xerox R&D, just of my $170 million for R&T. About 35% more of my budget goes to technology -- maturing what was in research before. The remainder is work in support of product programs.
We develop various platforms that Xerox exploits. A platform is not a product; it's a more generic concept -- a model that a business group can take and specialize for its particular markets. Of our product support work, about 85% usually goes to help our core businesses.
What's a core business for Xerox?
We identify three core markets. First is very large corporations. They need most of the products we make, and they are especially strong users of our high-end products: faster printers, color copiers, and multifunction office products. Second is the public sector. We are very strong there. And the third core area is the office -- not just the large enterprises but also the general office market. They each require different products. And to serve them efficiently, we need products that can be upgraded easily for the large enterprises. That's one function of the platforms we develop.
Do you also help translate the platforms into actual products?
Part of my job is to make sure, through the platforms, that technology gets shared among the groups, so that each group doesn't have to invent all the tech it needs. For example, it would be silly for every group to develop its own technology for connecting to a network. So we make a point of developing software modules that can be shared.
Does that mean you're reusing software?
Yeah, we do a lot of software reuse already. We recently created a database that has something like 20 million lines of code today. Eventually, every feature developed in software will be put in there, along with information about what it does. The level of reuse is not as high as we would like, but we estimate that savings based on reuse of code over the past two years was about $30 million.
You recently printed electronic circuits on plastic. Why work on that?
The flexible circuits technology could be a new business opportunity. But it also relates to our core business. Photoreceptors are a key component in Xerox machines, and the technology for making photoreceptors is evolving around polymer electronics, instead of semiconductors such as silicon. So we are developing a very strong competency in polymers.
PARC has been a nanotechnology pioneer. What's coming out of that work?
Until last year, all our toner was made by combining resins, pigments, and other ingredients to produce a big chunk of material that was then broken down into small particles. Today, at a pilot plant at our Canadian research center, we're growing toner from a molecule.
We start with a molecule that's a few nanometers across and grow it to a few microns in size. While we're doing that, we can chemically control everything, even form it the way we want -- a sphere isn't the best shape, because it's harder to clean. And since the particles are smaller and more uniform in size, the print quality is better. The particles get packed closer together, so the image is a bit sharper.
We constantly revamp our technologies, but after 40 or 50 years, we need very deep scientific understanding to go to the next level. Nanotech is a key to the future.