In the first of a two-part interview, William Weinberg, newly appointed architecture specialist at the Open Source Development Labs (OSDL), talks to vnunet.com about desktops, datacentres, security and the spectre of legal action by Microsoft and SCO.

We're seeing big-name companies joining OSDL. What are their expectations?

What the OSDL is most about today is the three initiatives we have established. The one that's best established is the carrier grade initiative. [Then there's] the data centre initiative and the newest one is the desktop initiative.

The goal of these initiatives is to limit the inhibitors to Linux adoption and to accelerate Linux adoption in the enterprise. Each initiative has a marketing working group and a technical working group.

What do you see as the general trend in Linux now?

In the desktop world there are divergent opinions of how the adoption should be tackled.

We have been involved in several attempts at building desktop systems to try to solve compatibility types of things like Lindows, work-alike open source office suites like OpenOffice, tools that let you actually run Microsoft applications on top of Linux, like CrossOver Office and Wine.

There's no one way to do it, but I think a very non-productive approach is to walk in the front office and say: "You have a Windows workstation today with these applications running and here's the short path to Linux tomorrow, with these other applications or the same applications."

Instead, the approach that looks like working in the enterprise setting [is] where Linux has done well in the datacentre and continues to advance on servers and infrastructure equipment, [so] the client device is an outgrowth of that adoption.

It is not a market of its own from our point of view. What we want to do is reduce inhibitors to the utilisation of client devices: desktop systems, thin clients, transaction terminals, maybe kiosks - the tools of the corporate information worker, the technical workstation user.

We want to make sure that, if they're part of a corporate infrastructure, those client devices are in a position to provide comparable or superior functionality to what they are providing today with a proprietary operating system, whether it's another Unix or Windows.

What do you see as the biggest immediate challenge?

Solving a lot of compatibility and migration issues in the datacentre: the datacentre is very complex. Companies have a mix of in-house and commercial software and they don't necessarily throw it all away. They want to have a migration path.

Open source has to be viewed in that context as a continuum, from pure free software à la GNU all the way out to wholly proprietary software that runs on Linux as a host and nothing more.

Being able to run the strategic killer apps in the datacentre without hiccups and without spending a lot of additional money on integration, and being able to migrate existing in-house applications, is a big challenge right now.

People are doing it, but everyone is having to pioneer that effort over and over again. Making that reproducible [as] almost an automated process is one of the challenges we're facing.

What about security?

Security has two big facets: actual security and the perception of security. What we don't want to see the world continue to do is rely on security through obscurity. We do not believe that is a good approach.

Openness is a great approach to security because all systems are hackable. People have been cracking proprietary systems for years without access to source code. The theory goes that source code makes it easier to perform certain kinds of cracking.

The big difference openness gives the world is that it gives [people] the opportunity to repair the fault and to do so in a more timely fashion to prevent further exploits, and so keep the system more secure.

And the community of folks who are interested in doing that is much larger, so we firmly believe that open systems over time are more secure than closed systems, and [that] at any given point in time they're more secure as well.

There have been reports of Microsoft using its patents against open source. What is your view on this?

Of the two I know about in a little detail, one of them is the Fat file system patent, and the other has to do with patents that are to do with codecs that get plugged into things like the Windows Media Player [WMP].

It's very hard to ship a Linux-based system that carries the WMP or compatible codec [because of] the restrictions Microsoft puts on licensing that code. But there's an odd twist in that a lot of individuals are allowed to make that integration themselves ad hoc after they have purchased the device.

On the Fat side, Microsoft has been attempting to play hardball. The information I have from the industry is [that] there is so much prior art and so many implementations of Fat-compatible file systems on different operating systems, Microsoft will have a hard time claiming they have ever enforced these patents - and thereby at this point in time restrict Linux use of that file system.

In the past OSDL has commented on the SCO legal cases against IBM et al. What do you think about the current situation?

OSDL and others view the current situation as more or less in stasis. SCO has had some problems in terms of its own investors looking at whether litigation is a good business model versus selling product.

Their lawyers have advised them to be relatively silent on a lot of issues and they seem to be taking that advice now as the actual trial dates approach.

What has been happening [is that] companies are less willing to talk about their adoption [of Linux]. But, while they're more reluctant to disclose, it's not actually slowing adoption.

We're putting money forward to defend AutoZone. Of the two so-called end-user lawsuits, only AutoZone actually involved Linux copyrights; at least, that's the way it was stated.

[The] DaimlerChrysler [lawsuit], like the IBM suit, appears to be primarily a contractual dispute, and therefore falls outside the realm of pure Linux IT litigation.

Companies like IBM and Hewlett Packard sell Windows as well as Linux. Is that an inhibitor at all?

These are [OSDL founder member] companies, and they obviously view Linux as sufficiently strategic to their business overall that they have invested in OSDL both purely monetarily - in terms of their funding efforts - and, probably as importantly, in the amount of time their executives and engineers give.

That doesn't mean they're doing it to the exclusion of other technologies like Windows or Mac OS. Look at IBM [making] such a large and public commitment to building on Linux: it doesn't mean they don't have Windows products.

Their notebooks ship with Windows and they're still shipping AIX and making money on it. It's just that they see Linux as a strategic lever for - in IBM's case - shipping their hardware in the server space, and shipping their silicon in the embedded space, and for creating opportunities for service revenues.

Part two of this interview can be read here.