Hewlett-Packard is associated with printers and computers, but its origins lie in scientific instrumentation. HP was also the first company to make a pocket-sized scientific calculator.
This month marks the 40th anniversary of HP Labs, its research arm. The European branch of the labs, in Bristol, is the American company's second-largest site and the main base for research into digital media and security technologies. HP also has research sites in India, Israel, China and Japan.
For HP to maintain such a large and broad-based operation requires a big commitment from its management and shareholders. Computer giants such as Dell maintain that they do not need to perform fundamental research, and other IT businesses have cut their spending in the area. Of its immediate rivals, only IBM has a larger R&D presence.
In its 2005 financial year, HP spent $3.5bn (£2bn) on research and development, out of a turnover of $86.7bn. In an industry where margins are thin, this is a big investment, though HP does not say how much of that budget is devoted to the labs.
"Along with most hi-tech companies, it has gradually reduced what it spends on R&D," says Andy Butler at the industry analyst Gartner. "But within the total funding, there is a difference between what the product divisions spend on R&D and what HP does in its labs. The labs are an autonomous organisation."
Over the past 40 years, HP Labs' inventions have changed not just HP but the whole IT industry. HP researchers took photocopier technology, licensed from Canon, and used it as the basis for the laser printer. In 1984, HP launched the first thermal inkjet printer.
The labs are also behind technology that lets companies buy computing power as a utility; technology that allows machines to communicate with each other; and technology for DNA analysis. They also helped make the movies Shrek 2 and Madagascar for the film studio Dreamworks.
For the HP Labs director Dick Lampman, working one step removed from HP's product teams gives his researchers the space they need to come up with their breakthrough innovations.
"The thinking behind the labs is that the intensity of pressure on any business unit means it is very difficult for them to fight their day-to-day battles in the market and set aside time to make investments in the future," he says. "It is pragmatic thinking. It is not, 'Let's make an investment in science' but about making sure we set aside resources to ensure we are successful both now and in the long term."
Measuring the value of such investments is complicated. Thousands of individuals might work on bringing a product such as the inkjet printer from a prototype on a lab bench to where it is today.
"Everyone struggles with measurement," Mr Lampman says. "In modern business, that's where the focus is. But we are not trying to go through a spreadsheet exercise. In research, you would have to make so many assumptions that you could produce almost any numbers you wanted. But in qualitative terms, managers [within HP] know if we are having a big effect or not."
A degree of autonomy is vital for HP Labs. Sometimes, researchers can find business units are sceptical about the value of a piece of research. Sometimes, R&D programmes meet with opposition.
This is something Mr Lampman does not just accept but feels is essential for good research. "Sometimes we invest in things there is not a lot of support for, and we are very transparent about that," he says.
"The most spectacular example was the inkjet: at the time it was counter-strategic for HP and There was legitimate logic behind that position. But we believed there was room for two mass-market technologies. It was a good call."
Good science is also as much about eliminating lines of enquiry as the inventions that do make it to production. For every "eureka" moment there will be dozens of ideas that do not work, that work but are not financially viable, or are just ahead of their time.
"This is something you don't see on the honour roll, but it is part of the total system that makes us work," Mr Lampman says. "We have had two print technologies for a long time, but we have been looking at LEP [light-emitting polymers] since the early 1990s. We need to understand alternatives that might open up new spaces or displace existing technologies. That is as much part of our responsibility as looking for the silver bullet that might change the industry again."
Nor is HP Labs afraid of making strong statements about the direction of its work. One research theme has the bold title of "re-inventing the economics of IT", focused on massive data centres where HP could offer computing power to companies by the process or by the second.
HP also has a research team working on quantum computing, especially in semiconductor design. Mr Lampman believes nanotechnology research will be especially important for HP in the future, for example, in enabling chip designers to put many more circuits on to a device than today's technology allows.
There had been fears that Mark Hurd, appointed chief executive last year after the ousting of Carly Fiorina, would cut back on the labs as part of taking HP back to basics. In fact, his review of the labs found that the research corresponded to his vision for HP. "We are supposed to be ahead of the road map," Mr Lampman says. "That is what he likes to see."
THE INVENTIONS OF HP LABS
1966: HP introduces its first computer, the 2116A.
1967: HP atomic clocks are used to set world time standards.
1968: HP Labs develops the first commercial LEDs.
1972: HP Labs produces the first pocket scientific calculator.
1980: A 64-channel ultrasound scanner is developed based on HP technology. HP also creates the first commercial office laser printer.
1984: HP launches its first thermal inkjet printer, based on research done since the 1970s.
1991: HP Labs develops acceSS7, a probe system that taps into the global telecoms network, cutting fraud and improving performance and billing.
1994: HP joins with Intel to develop the Itanium range of microprocessors, based on work started in 1981. Itanium launches in 2001.
1999: HP and university scientists build a molecular logic gate, which is a step towards chemical assembly of nano-components.
2001: HP Labs Bristol develops active counter-measures to combat the Code Red virus.
2003: HP announces services based on grid computing.
2004: HP compression technology carries images of Mars to Earth from Nasa's Spirit Rover craft.
2005: HP Labs helps produce Shrek 2 and Madagascar. Researchers prove that an HP Labs invention, the crossbar latch, could replace the transistor in building computers.
And in the future...
Utility computing: HP will combine large numbers of industry-standard computers in huge data centres. These will be cheap to run and require little human intervention. They will pave the way for companies to buy computing resources the way they buy gas or electricity.
Visual radio: HP researchers in Bristol have developed a technology that combines digital radio broadcasts with the internet, for transmission to portable sets or, potentially, mobile phones. The radio's screen can carry news and traffic updates or links to services like music downloads.
Digital paper: Also in Bristol, researchers are working on printable plastic displays. As thin and portable as paper, they could change advertising hoardings and, ultimately, newspapers.Reuse content