The man guiding Oracle Corp.'s Network Computer, Inc. (NCI) subsidiary day-to-day is CEO Jerry Baker, who took the job a year ago at the behest of his boss, Oracle Chairman Larry Ellison. Baker recently shared his views on the network computing market with Network World Senior Editor John Cox.

Do you really believe the Wintel platform is a dinosaur?

I wouldn't include Intel with the dinosaurs. It's not the Intel platform [that's outdated], but the software that you install and have to manage.Wintel is just not a workable system for most people.

And what is?

A server-based model where you pull information from the server as you need it - applications, data, content and services. That's the model most appropriate for both the consumer and corporate environments.

There's a massive investment in Windows software. Surely people aren't going to scrap all that.

The same objection could have been made 20 years ago with the advent of the PC: There was a huge investment in IBM mainframes and software. But the PC offered such compelling benefits that a whole new model grew up around it. If you look at investment being made right now in Java - and Java is far from perfect - the resources going into developing Java applications is incredible. And the applications will all run on the network computer (NC).

Are you going to make money with this?

Our goal is to generate revenue [through software sales], become profitable and eventually take the company public.

Are you profitable now?

We hope to be in 12 to 18 months. We are selling our software now but in very limited quantities for two reasons. One is we're a small company and we have limited ability to support customers. The second is the manufacturing for the NC devices is still ramping up.

How do you plan to create NC interoperability at the server level among NC product suppliers?

In three steps. First, at the applications level, so that an HTML or Java application on an NCI NC will also run on a Sun JavaStation or an IBM Network Station [or others]. Second, through the smart card [a microprocessor device the size of a credit card that holds the user's data and personal configuration information], which if based on the standard we announced a few weeks ago, will work in any NCI-based NC. Third, being able to run my NCI NC off an IBM server. This level of interoperability is being discussed and worked on right now. I hope by the end of the calendar year to have an architecture and specification to implement that.

What are NCI's goals for the rest of '97?

In the corporate marketplace, to create a stable portfolio of reference account customers who will deploy NCs in meaningful applications. We want 10 or 12 of those this year. In the consumer market, I'm not looking for huge volumes, but rather for validation that the consumer NC market exists.

Source: Network World