Interview by David Needle

How is this for a business plan? Go out and develop application software in an area where the industry leader has over 90 percent market share, a loyal customer base, and billions of dollars to spend on R&D. It may sound crazy, but that's just what Ken Rhie and his co-founder (Chief Technology Officer TJ Kang) set out to do three years ago when they started ThinkFree in Cupertino. Since then the company has grown to 120 employees and without using any advertising, says it has signed up close to 300,000 people to use its free software just emerging from an eight-month beta period.

ThinkFree has developed a suite of Microsoft Office-compatible applications for the Internet, written in Java, called the ThinkFree Office. But rather than try and woo users away from the Microsoft Office virtual monopoly, as Lotus and Corel have failed to do after years of trying, ThinkFree Office hopes to be a welcome compliment for Office users that they can access from any Internet-connected computer, and use offline as well.

The ThinkFree Office has been free since its release earlier this year, and you can still get a free trial version for a month of use. An always-available version without onscreen advertising costs $24.95 per year, which includes free updates over the Internet and 40 megabytes of free online storage for data files (the trial version comes with 20 megabytes).

Why ThinkFree Office? What was the opportunity you saw three years ago?
At a high-level perspective, computing paradigms change about every decade, when a different way of communicating emerges. I started with mainframes, then minis came along, then PCs and then the graphical user interface. Windows dominated for 10 years. Now I believe another paradigm migration has begun that's focused more on network enablement. I believe every time there is a shift like this you need to redesign killer apps from the ground up. For example, we had editors for mainframes, but we needed WordStar for the first personal computers. The architecture changes but the need to do these applications doesn't.

Of course you're not alone in hopping on this network computing bandwagon.
Sure, even Microsoft recognizes this with its dot Net initiative. But all there is right now is vaporware and words; there's no actual implementation. We started on this three years ago and called it Web top software and now we've done it. We give you word processing and spreadsheets that look and work [like Microsoft's] but leverage the Internet, and network computing. We're designed for the new networked environment, while Microsoft is adapting to it.

Sun offers its StarOffice for free; aren't they a competitor?
What they're doing isn't architectured for the Internet. You have to download 180 megabytes of code-that's an office all right. Ours is less than 10 megabytes of code and you can download it over a broadband connection in a few minutes, and once you're registered it's available from any PC you log into. A compact size like ours is more conducive to the Net.

You're going to make a big announcement at Comdex?
Yes, it's a major ASP [Application Service Provider] I can't pre-announce, but they are a major brand with millions of users, which will give us great exposure and credibility.

Sun recently made the code behind its StarOffice software suite-presumably a Microsoft Office competitor-available to the public as open source. What do you think of that, and is it something you would consider doing for ThinkFree?
Not at all for us. To be effective with open source you have to assume there are thousands of people who can or want to contribute to the cause.

Like Linux?
Exactly. There are all these Unix experts who know Unix code inside and out so it sort of makes sense they would be attracted to tinkering with Linux [which is a variant of Unix]. But for word processing or spreadsheets how many people have a background in the code of those applications? How many majored in word processors? Our code was entirely developed by ThinkFree over three years so there is no expertise out there in our code.

So why do you think Sun is doing it?
They are giving it away because it's not that valuable to them. They have no interest in making money with [Star- Office], it's really just "annoyware."

Annoyware?
Right, they're just doing it to bother Bill Gates and Microsoft.

Can you give us some projections as to how you will evaluate your success say a year from now?
The Net migration isn't happening today in big numbers; it's still considered cutting edge and in the evaluation mode. You won't see 10 million users next year, but even Microsoft sees the trend coming. I haven't talked to anyone in the industry who does'n say [Internet applications] are the right thing to do. We have several hundred thousand customers today who signed up for free, and a year from now we will have gained a lot of paid customers, tens of thousands. Also, a year from now we will have released additional benefits we haven't shown yet. We'll add another dimension that gives you more value-add on top of the Office suite.

What is this added dimension?
It's nothing similar to what's out there now, it's a new capability. ThinkFree Office was a huge accomplishment, but I'm not satisfied with offering basic components like the engine and tires. This will get the car running with a new dimension of benefit, but that's all I can say for now.

You could do quite well with just a tiny percentage of that. Is that your goal?
In three years there are projections that MS Office could be bringing in as much as $11 billion, so 1 percent, or over $100 million is plenty of money, but you don't see many survivors with 1 percent market share. A better question is "Can I grab the lionshare of this emerging segment of Web-based Office applications?" And the answer is, "Yes." It's just like Red Hat is a leader in the Linux market, but a tiny percentage of the desktop software market. We are a leader in this emerging Web-based applications market and most of the others are hot air announcements.

You think a segment of MS Office users will want to use ThinkFree as well?
If you have Microsoft Office on your desktop PC, the minute you walk away from your own computer to another office, or to another city, you want accessibility. You can use ThinkFree as a service that complements MS Office, just as Hotmail complements Outlook.

Since you don't have CDs and a physical distribution network to worry about, what does it actually cost to supply a new user with ThinkFree Office?
It doesn't cost us a lot, just extra bandwidth for the traffic. And that com- pares to Microsoft, which has to look at a Web-based Office as cannibalizing its CD sales. For them, it's a lost opportunity to pocket $200 to $300. My costs are simply out-of-pocket. The exact cost depends on how many users we sign up but it's definitely under $10 each.

You have another product in the works called OfficeMate.
We believe you should be able to access information from anywhere, even if you have a dumb device with a low clock speed, whether it's a Palm-type handheld or even a video game player. OfficeMate is lighter client software that will run on smaller, non-computer devices that can't run full-blown Java applications but can connect to Internet servers and retrieve documents. It's not finished yet. I'm a firm believer in credibility, so I'm telling you this as a peek into what we're working on.

You've bet big on Java. What's your experience with it been like? Is it living up to its "write once, run anywhere" promise?
If you know how to design an application using Java in the first place, write once, run anywhere works. ThinkFree works on Windows and Linux. We did not have the expertise to do this from the start, but we learned, and it took a lot of trial and error to get right. Is it as easy as Sun's marketing literature says? No. But it's way easier than trying to do the same thing in C where you have to rewrite everything. Relatively speaking, Java is extremely portable where C or C++ isn't.

What about ThinkFree for the Mac?
Later in November we will be shipping a version for the Mac through an OEM distribution arrangement in Asia. That code is close to what we did for Windows and Linux but slightly different because of the Mac UI [user interface].

Are there plans for a Mac version for the U.S. market?
Well, Microsoft has not said when or if they will port Office to Mac OS X, so we may be able to make Thinkfree Office available before Microsoft. If Apple doesn't have Microsoft on board that will definitely impact the launch of OS X. I would say to Apple that Microsoft is not the only game in town.

You have another product in the works called Oasis ÷
Again, this is a future direction, and not something we have available today. But later in 2001, Oasis will be our solution to provide the entire ThinkFree Office solution to corporations so they can host it within their own firewall at their own company.

David Needle is managing editor of the Web news site SiliconValley.internet.com. Contact him at dneedle@internet.com.