by Nancy Cohen

Transvirtual may become a very notable player, as a company set to create a new communications standard for Internet devices. Founded in 1997 by Tim Wilkinson and Peter Mehlitz, last year Transvirtual got generous trade press attention as ‘that company with PocketLinux.’ This Open Source framework application platform aimed at embedded systems is sold by Transvirtual through their direct sales force, VARs, and system integrators.

What that offers is a fast, cheap solution for resource-constrained information appliances such as PDAs, web-enabled mobile phones, automotive telematics, and TV set-tops. What it gives Transvirtual: Good business.

XOE is the commercial implementation of PocketLinux. The technology is covered by patents. XOE is licensed to OEMs, while much of XOE is released as Open Source under the GPL. XOE will be marketed by Transvirtual to embedded system and device manufacturers. Mehlitz tells Open why XOE will make a real difference for developers servicing corporations about to turn their corners toward web environments.

Q. Kaffe is said to be a clean-room effort. What’s the definition of ‘clean room,’ and what should developers know first and foremost?

MEHLITZ: 'Clean room' means the Java implementation was built from scratch by using public Java specifications. This does not impose any restrictions on how the system is used. It does get interesting if this is about the development of an Open-Sourced and clean room Java implementation itself. In this case, you have to ensure that no contributor has been exposed to sources that spoil the clean-room status of the project. Many developers don't even realize they get tainted by this, and subsequently can no longer participate in Open-Sourced clean-room Java implementation projects

Q. So what do they need to know?

MEHLITZ: The way to prevent losing clean-room status is to use source repositories that have restricted-write access, making sure that everyone who can write to it has signed the legal paperwork that he or she has not seen other Java implementation code. [The Kaffe developer site explicitly states that one cannot contribute to Kaffe if under a non-disclosure agreement with Sun.]

Q. For programmers who have been working with Enterprise Java Beans, how will an XML-based XOE platform make a difference, in terms of the ease and powerful stuff that they can build with?

MEHLITZ: The best way to answer that is to say what XOE is, which goes far beyond a XML-related Java class library. XOE is a whole software architecture. It consists of client and server side, with the sole purpose to bring as many different types of content to as many different devices as possible.

Every XOE layer is based on standards and well known formats, protocols and APIs (XML/DOM, SOAP, Java, Posix). Hardware-specific functions (e.g., drivers or video players) can be brought in at the OS level. The programming environment can be used to run standard Java applets. The application framework can handle standard HTML, as well as application specific XML. External services are accessed via SOAP. There is no risk of being restricted to a subset of applications and devices, or to bet on new formats, protocols, and languages which are not yet accepted by the market.

Q. Who does Transvirtual see as its major competition?

MEHLITZ: One can find competitors at different Transvirtual technology layers, such as Java implementations, web browsers, and PDA application suites. But there’s no direct competition yet in providing the whole spectrum in an equally generic and extensible way. This is important news for vendors who need to support multiple-device categories and don't want to be entangled in the web of software integration.

Also, Transvirtual is not a competitor in the operating system business. Transvirtual cooperates with OS vendors. On the server side, using XML/SOAP is a good basis for interoperability with big vendors like IBM and Microsoft.

Q. What do you see as important new technologies in driving web services?

MEHLITZ: Wireless networking will certainly play a significant role. Even though XOE does not directly depend on a specific physical network, 2.5 and 3G technologies will boost applications and devices that are targeted by XOE.

In the embedded systems market, Linux will gain further momentum, and it is just waiting for an application framework enabling widespread use in end-user devices. As for upper XOE layers: Java, XML, and SOAP are already established standards for heterogeneous networks.

Q. One of your press releases says that "Manufacturers, OEMs, Systems Integrators, and Enterprise Developers will see immediate benefit when leveraging XOE for platform-independent application development and distribution." Why?

MEHLITZ: The application developer sees benefits because XOE does not try to re-invent the wheel. XOE enables new applications by combining and extending already known technologies. Manufacturers and OEMs benefit by having a single, portable software architecture that enables them to easily integrate device-specific functions and brand their devices with a corporate-specific look and feel. Integrators don’t have to deal with incompatible or redundant components

Q. I noticed that a non-HP site news story about HP’s Cooltown project explained the corporate vision for “a world filled with interconnected computing gadgets.” I wonder what you think about references to "gadgetry." Would you see the next wave of web services for content and application delivery as something far beyond gadgetry?

MEHLITZ: It's not about silly gadgets; it's about extending the functionality of existing products by means of Internet connectivity. One aspect is re-using the same content on different devices (e.g., the address book on your PDA, cell phone, laptop and desktop systems). Classical PC 'syncing' isn't really an option if more than two devices are involved. Content distribution is also important for potential pay-per-use systems like set-top boxes and gaming consoles. Telematics systems will combine entertainment, safety monitoring, communication, navigation, and up-to-date traffic information.

Q: Back to the real-world frontlines of corporate computing: When do you see more corporations actually ready—that is, with the kinds of computing strategies that can actually incorporate the XOE architecture and web service applications? Are we seeing the first wave of adoption?

MEHLITZ: Adoption depends on the availability of wireless networking. Connectivity is a crucial factor for many XOE functions, which makes the Japanese market especially suitable for early XOE adoption.

Rich content-enabled, networking capable devices are already in development, and will be ready for deployment by the middle of next year. XOE will be able to show its strengths on these devices.

Linux for end-user devices is also fairly new, but recently has gained tremendous interest, especially from device manufacturers. First products are already available, and a lot of Linux development effort is now focused on this.

Security will play an important role in corporate applications. Technologies like SSL and VPNs are already available and can be used efficiently on the next device generation.

To answer your question directly, widespread adoption in corporate computing will certainly take its time. This is not something that grows out of unproven technologies or rushed test phases. But it will happen. We should not think of pervasive computing as something that solves problems we don't have yet. It will be much more like e-mail, making a lot of things in our day-to-day life easier.

The common denominator is the extensive use of XML. We expect 2002 to be a good year for the first major deployments.