Linux and the open source movement is rarely out of the spotlight these days. And the debate about its credibility as a viable alternative to Microsoft garners much attention.

Red Hat is the latest Linux company to challenge Mircosoft's hold on the business desktop space with the introduction of Red Hat Desktop, the latest addition to the Red Hat Enterprise Linux family.

Matthew Szulik, the company's chief executive, chairman and president met Computing to discuss Red Hat's latest efforts and share his views about the Linux landscpae.

Why has Red Hat now decided to release a corporate desktop product?

I speak to very high-level customers and government leaders who are starting to build next generation information systems. It starts with building on standardised hardware. We're starting to see the [necessary] application performance on Linux.

With dependable architecture, infrastructure and security, we believe we can now achieve what they need on a desktop - not flashy 3D graphics or a cool game. The hardware community has been writing device drivers for one supplier for 20 years. It has to be made aware that there's a burgeoning Linux client market before they start moving their device interfaces, drivers, adapters. Because of Red Hat's success in enterprise computing in the last 24 months we're seeing them starting to make Linux drivers.

Red Hat wants to solve the problem of allowing data and infrastructure to be managed securely and reliably. The biggest concern for the IT guys is security. What ports have been left open installing another laptop? When adding another system user, does the security infrastructure scale reliably? Are they compromising any protocols as more people join the network?

You're seeing companies like Salesforce.com emerge that will deliver that sales force automation functionality as a managed service in the same way we're distributing our desktop.

Why do you think a Linux desktop will succeed despite Windows having a 96 per cent market share?

The computing paradigm is changing. Whether it's your word processor, email package or calendar, it will be a web service in five years time. The desktop was built over twenty years. Most people in 1995 wouldn't have thought email would be their preferred method of communicating with their loved ones.

How in such a short time has Windows gained exponentially more applications running on it than Apple? Part of it was clarity of focus. Macs got into the server way too late. Red Hat started out [in 1997] and stayed focused on the back-end server, because we believed the network was the future of the industry. Apple and Microsoft to a large extent continued supporting their proprietary franchise on the client, and arguably missed a very large part of the server market.

Why Red Hat versus Novell-SuSE or Sun JDS on the desktop?

They're all proprietary except us. They all have proprietary technology inside, not 100 per cent open source software. They continue to lock customers in to limit choice.

If you buy the Sun desktop, you're going to buy into the proprietary Sun architecture. With SuSE there's Red Carpet, integrated with other Novell technologies, still proprietary.

If you're a Microsoft customer and your options continue to be proprietary software A or B plus open source, why change? It's all this incremental functionality that customers have got tired of spending money on for the last 10 years.

Will Linux applications mostly be written in Java in the future, and will Linux fragment?

It's undecided. Work is ongoing on an alternative to Java or .Net, essentially an open source Java implementation. There's substantial work on Jonas [open source J2EE application server, now Red Hat-supported] through the non-profit consortia.

Comparison to Unix will always be an issue. But the reality is, if the source code is in the public domain, you can take whatever kernel, whatever libraries, you want. The key issue is application support. Which kernel in which Linux will all your major ISVs choose to certify on?

Is .Net on Linux a viable option for the future?

.Net has taken so many twists and turns. It's an incredibly bold vision, but implementing it is another matter. The challenge for open source is: can it continue to scale as fast as we do now, to provide a technical alternative? I don't think there's much doubt about that. Producing an operating system is very expensive and very hard. It isn't the functionality that becomes the differentiator, it's one proprietary development model versus an open source collaborative development model.

It's no different from the assembly line of Henry Ford making automobiles versus the modern pre-production assembly techniques being used by the Japanese today. It dramatically changes the economics.

How do you feel about Linux web services?

Web services is how we'll distribute this desktop product. If you go to redhat.com or use the Red Hat network, that's a web service [using] XML-RPC. The issue about web services is vendor neutrality if one vendor controls XML, or chooses to license the XML format or XML data schema.

What is your take on the SCO situation at the moment? What impact has it had on Linux?

Lawyers continue to make money. We sued SCO. We're waiting on the judge. She said she was going to hear the IBM case first.

We're starting to see industry consortia get together. Large Wall Street banks are making large configurations of Linux. They get together. They're getting value from it; they're sharing ideas.

These Wall Street firms are getting great economics from standardisation on Linux and Intel. They're willing to talk to each other about their technology and how they're deploying it, because they know the advantage is not in the technology; it's in their business processes. This is something many customers have missed over the last 20 years. Take the Harvard Business School article, 'IT doesn't matter' by Nicholas Carr. His point is if you and I use the same technology, the differentiator is not the bits, it's how we deploy it.

What do they these companies see as the Linux advantage?

Cost, performance, standardisation and no vendor lock-in. They're the big four. Lower operating cost, improved performance, improved reliability and not tied to a single vendor. They were saving [millions] getting rid unreliable systems causing lots of expense and headache. One of those Wall Street banks now has one administrator for 800 machines. One did it then everybody else came rushing to him to say: 'how did you do that?' Now nine out of the 10 leading Wall Street banks are Red Hat customers.

I notice, because of the transparency of Linux and open source, that the software development community being tasked with building these mission critical systems are not buying into the marketing hype and the vendor-sponsored reports from analysts.