Father of Java Has His Eye on Jackpot

Interview: James Gosling talks about Web services, open source, Java (of course) and his new project 'Jackpot'.

In the seven years since its introduction, Java has made rapid advancements into the enterprise. In the burgeoning field of Web services, Sun Microsystems Inc.s Java-based technology is in tight competition with Microsoft Corp.s .Net initiative for the heart and soul of developers. James Gosling, a Sun vice president and fellow, and father of Java, spoke with eWEEK Senior Writer Darryl K. Taft over the summer from Goslings Palo Alto, Calif., office about Web services, the future of Java, open source software and its impact on the software business, and Suns success in the tools business.

eWEEK: I want to talk a little bit about whats close to you, what youve been dealing with in the lab. Can you talk about what you are working on?

Gosling: A little bit. Its a research project. Theres not much reality there, but Ive been looking into issues around developer tools, working with the NetBeans group. And Ive been mostly working on analysis and transformation tools based on having a complete semantic model of the application.

eWEEk: What do you mean by that?

Gosling: Its one where I have the application as a database and then can do analysis on it, though its not exactly a database, its more data structure.

I keep sort of an annotated parse tree, which means that instead of the way that most tools look at programs as a series of lines and text, with punctuation and letters, left to right on a page, top to bottom, I actually have all of the different entities all related so I can do things like find all the places a particular variable was used, trivially. If I want to rename a class, thats a trivial operation. Youre including accounting for changing that name every place that its occurred. Its easy for me to do things like if youve got any such variable and you make it private, then go and find all the places where that variable is used, turns them into instantiations of access or methods, and if the access or methods dont exist then construct them as well. Thats not a difficult thing to do in my experiment test bed.

Its still kind of early. Its still a research labs project. The project is called Jackpot.