John Ousterhout is a proud father. He single-handedly developed Tool Command Language, or Tcl, a scripting language which “glues” together applications written in standard system programming languages such as C, C++ and Java. As a professor at UC Berkeley, Ousterhout (pronounced OH-stir-howt) wrote this new technology in the late 1980s as a solution for his own programming needs, then began letting others in on his secret. Its popularity has since soared. After briefly planting the Tcl movement at Sun Microsystems, Ousterhout is happy to announce his latest offspring, Scriptics Corporation, a company formed around the business potential of Tcl.
While on staff at UC Berkeley for several years, Ousterhout handled the typical professorial duties of teaching, researching and delivering papers. Little did he know that he would stumble upon an idea which would form the basis of his career. Revealing that he “got into scripting by total accident,” Ousterhout recounts that he had been working on computer design tools, each of which needed its own command line language. Never having enough time to write a comprehensive solution, Ousterhout would throw together a different piecemeal system each time. Fed up with the sloppiness of the situation, Ousterhout built the core of a language that could be used over and over again with various applications. This was the basis of Tcl. “I kept building bad language after bad language,” he says. “Tcl really grew out of embarrassment.”
Tcl was Ousterhout’s baby. As a professor, he often let the graduate students handle the fun development end of projects and relegated himself to handle the grunt work. Happy to find a new creative outlet, Ousterhout kept Tcl all to himself. That is, until he started handing out free copies and, like most great software, the language’s popularity began to spread through word of mouth. First distributed in 1989, by 1993, Tcl had garnered tens of thousands of users and today boasts more than a half a million developers. This grassroots movement was something Ousterhout never predicted, but he immediately recognized its potential from a business perspective. After 14 years in academia at Berkeley, Ousterhout chose to follow his passion for his pet project and, in 1994, accepted an offer from Sun Microsystems’ labs to build their Tcl initiative. Having consulted for Sun in the past, Ousterhout was confident in their ability to provide him with the resources he needed to make Tcl the definitive scripting language for the Internet.
The transition from an academic lab to a corporate one was smooth given the similar small-team atmosphere — a boon to research. “Everyone used to ask me if I experienced culture shock, but I never did. The only difference was that I wasn’t teaching anymore. A lot of great things happened at Sun.” These great things included the porting of Tcl to the PC and Macintosh platforms, which brought an explosion of new users.
Ousterhout’s satisfaction with Tcl’s standing at Sun soon hit a major roadblock, namely Java. Just as Tcl was gaining serious momentum, Java’s popularity and the ensuing war with Microsoft bubbled up to the forefront of the entire computing industry. It became obvious that Sun would need to throw all of its weight behind promoting and defending its creation, thus leaving little room for Tcl on its agenda. Noting the shift in focus, Ousterhout decided earlier this year that it was best to break from the giant Sun and start his own Tcl-centric company, now called Scriptics Corporation. Claiming the split was very amicable, “Sun even gave us our first Web server,” he says, Ousterhout is now faced again with an entirely new working environment as CEO of a start-up.
“Tcl had been very successful, but it made sense for Sun to put all of its resources behind Java. There just wasn’t energy left to expend on Tcl and, finally, it made sense to split. On the list of things I’d like to do in my lifetime, doing a start-up had always been there.”
Though he originally intended on filling the role of CTO at the new company which, for the short term, is focused on catering to the needs of the established base of Tcl developers, Ousterhout quickly assumed the CEO position. “I used to think I had to write 10,000 lines of code a year to maintain my sanity. Now, I like being in a position where I have a tremendous variety of tasks.” He welcomes this new role, but realizes that, as the company grows, a more seasoned executive may be brought in to run the business. Even so, this opportunity feeds both sides of Ousterhout’s working persona. “If I had to characterize myself, I’m sort of a researcher and a practical person. To put it in a sentence, what I like to do is create innovative software systems that change the way people build and use computers. I like coming up with new concepts, but the real thrill for me is to see people be more productive using the systems I developed for them. Tcl has allowed me to do that.”
In discussing the benefits of Tcl, Ousterhout takes a much broader look at the software industry. He describes a fundamental shift happening now in the way people program. “We’re moving to a two-phase style,” he explains, “where you build components in C, C++ or Java, then glue these components together with a scripting language to integrate the tasks. This new style of applications is becoming more and more important.” Ousterhout’s own interest in founding Scriptics is to be the purveyor of such a development model. “This shift is beginning to enter the mainstream where Fortune 500 companies are using scripting to build their systems. Our long term goal is to complete that transition.”
Tcl’s success story has been encouraging for the open source software movement, which has been making headlines lately after receiving Netscape’s support. Ousterhout describes his own role in the effort as “actively involved, but not a religious fanatic.” He acknowledges that in distributing the Tcl software for free he was able to develop a dedicated group of developers, but disagrees with those who “believe that software ownership is fundamentally evil.” Describing open source software as “a fabulous empowering device,” he explains, “In a world where it seems like more and more things are controlled by large corporations, open source software makes it possible for a single person with a great idea to change the world and harness the energy of thousands of other people.”
Ousterhout’s introduction to computers came in the 1970’s while he was an undergraduate at Yale. As a physics major, he learned FORTRAN and was hooked. “Half-way through junior year as quantum physics equations were running from one page to another, I realized that the only thing I liked about physics was playing with the computers in the lab. I’ve been a computer person ever since.” Though it was too late to switch majors at Yale, Ousterhout decided to enroll in a computer science PhD program at Carnegie Mellon immediately after graduation. In 1980, Ousterhout headed west to the University of California at Berkeley to join the computer science department as a junior professor, where, true to its reputation, Ousterhout was impressed by the friendly atmosphere. “Some places are really competitive and tend to assault young faculty and if you survive, you get promoted. Junior professors are really nourished at Berkeley.”
Having spent almost the past two decades in Silicon Valley, Ousterhout is amazed at the current landscape of the computing industry. “The whole economy here is like a primordial evolutionary soup where new genetic strains are coming into existence every day. All of a sudden there’s a company that’s the equivalent of a six-headed monster which goes ballistic for three months, and then the environment changes and this creature is no longer favored and some other creature comes to the forefront.” Whether Ousterhout’s own Scriptics evolves into such a beast remains to be seen.
Please send any comments to Amy Oringel at: [email protected].