James Gosling, the father of Java, is now the chief technology officer of Sun Microsystems's developer products group. In the aftermath of the company's big JavaOne event this week, he weighed in on the company's difficulties in the commercial tools space in an interview with Computerworld.
Sun puts on this annual JavaOne showcase, yet few developers use Sun's tools. How do you feel about that?
James Gosling - I think up until recently our tools haven't been as interesting as they should have been. With NetBeans 3.6, we finally have a release that has very crisp performance and a really good feature set. Until that came out, we haven't been able to push it as much as we would like. NetBeans has been out as an open-source project for, like, four years, but it wasn't until a year and a half ago that it really got energy put behind it. We've been getting really good numbers of downloads of NetBeans. And Creator is really the first product that we've launched that has broad market appeal, because it is targeted at the Visual Basic developer -- the sort of person who doesn't want to spend all of his time thinking about the guts of writing code. They need to be able to build these things very rapidly.
So do you think Creator and NetBeans 3.6 will start to change the situation?
Yeah. And the 4.0 release of NetBeans is very close to being ready. And it's lovely.
What's so great in the new version of NetBeans?
We've put a lot of effort into cleaning up the user experience, making the user experience less geeky, making the workflow a lot smoother. We've had huge, huge pushes on performance to make it very sprightly. The NetBeans team tended to be very focused on academic purity. Getting them to be a little more blue collar was a challenge. But they've really gotten the religion. And now they've put a lot of effort into polishing the details.
What's in store for the future of Java Studio Creator?
You can drag and drop components and build these apps beautifully, and subsequent versions of Creator are going to broaden that scope. The next version of Creator, for instance, isn't going to be limited to generating HTML as the output but [will] be able to build rich clients. We'd been hoping to be able to put down all that this year, but it wasn't quite ready.
When will it be ready?
Hard to tell. Certainly next year is kind of the hope.
What has changed at Sun that is causing this new push in tools. Was it Jonathan Schwartz deciding that he wants to have a tools business?
There was a combination of things. Some of it was with Jonathan. I mean Jonathan has just been a breath of fresh air all over the place. Part of it was, if you wind back the clock five or six years, that we in general tried to foster an external ISV community of people doing interesting tools to doing stuff of all kinds. And five or six years ago, the fantasy for us was that there were companies out there doing tools. But the reality of the tool business is that the economic models are just horrible.
Microsoft set the list price of tools at something so low that it's almost impossible for an independent company that's just focused on tools to actually survive. And so you've seen this incredible winnowing of the tool market. I mean, there's essentially only one left. There's just Borland. And it's pretty shaky. And the developer community has gotten addicted to this world where they spend less on tools than they spend on lattes.
Isn't the increasing usage of Eclipse and NetBeans also having a big impact on the tools market?
We open-sourced NetBeans a year before the Eclipse project started. And they [IBM] keep asking why don't we join Eclipse. And it's like, why didn't you guys join NetBeans?
Eclipse does appear to have garnered more momentum.
IBM has put just tremendous funding behind the marketing push for Eclipse at a level that is pretty impressive. But I think certainly on the technology side, if you look at the quality of the products, we are probably actually better and more complete than Eclipse is today.
Which version of NetBeans, because some developers say NetBeans is slow and clunky?
Most people who say that haven't actually tried any recent version of NetBeans. If you look at the old versions of NetBeans from two years ago, it really did seriously have performance problems. But if you look at today's NetBeans, it's dramatically better. I mean, it's just night-and-day better.
So, you think much of the problem is perception?
Our No. 1 problem right now is perception. And we're working very hard on dealing with that perception problem.
But how do you make money if so many people say the free tools are good enough?
There actually is a market where people are paying, and that's usually the more specialized tools. And the way that NetBeans is architected, it's not so much a tool as a tool environment. We've all released all these little higher-level tools that are built as plug-ins to it. So Creator is really just an extension to NetBeans, and Java Studio Enterprise is an extension to NetBeans. The mobility stack is an extension to NetBeans.... The things we've seen so far are pretty encouraging, so you can actually turn it into a sustaining business.
Is it your charge to make a lot of money on tools?
For us, the real goal is to make it so that the software ecosystem is as healthy as possible. We don't have to be fabulously cash-positive. We just have to be not fabulously cash-negative, because these tools are enormous. They're tremendous engineering efforts. There are a lot of tools companies that are like three guys in a garage. In the tool world, you just can't compete with a team that small.
Still, despite the fact that Sun created Java, its application server and your commercial tools remain at the bottom of the market-share list.
You've picked a couple that have been in trouble. The app server is doing a lot better. Recent market share data for NetBeans is noticeably better. The app server thing for us was just a Marx Brothers comedy. There was this period where you couldn't buy a company without getting another app server, and then this huge political mess (followed). We literally had like six (application servers).
A mess?
It was a huge mess.