Eclipse dreams up new schemes of disruption

"[Eclipse] has definitely changed their business because it isn't enough to just provide the basics anymore. You can get basics from open source, like Eclipse. If you look at the products that are unique and differentiated from each other, it's on feature sets. There's room to build innovative projects on the platform," Milinkovich said.

Looking ahead, RCP is of major interest to the group. Eclipse is waging a PR war to convince Microsoft developers they should use RCP as it provides backwards compatibility between Windows Vista, due in January, and legacy clients like Windows XP. RCP also helps small ISVs save R&D dollars while broadening their addressable market because RCP applications can be compiled to run on Linux and Mac - not just Windows.

Eclipse cites an Evans Data Corp poll saying 44 per cent of developers have committed to use RCP in the next six months, with "the bulk of respondents in enterprise IT organizations with over 1,000 developers."

"RCP solves a hard problem with style and grace," according to Milinkovich. "It provides a common programming model that gives developers the ability to build rich desktop applications spanning multiple platforms and to integrate those platforms with a mechanism for deployment and management."

The future is not without its challenges, though, and growth is clearly breeding issues. The proliferation of projects and plug ins is putting Eclipse under strain, as developers complain that using too many plug-ins in an Eclipse IDE kills performance levels. Also, it's becoming difficult to see the IDE wood for the plug-in trees.

Eclipse hopes to remedy this with a central web site and rankings designed to help developers find the plug-ins they want. Project Callisto, meanwhile, tries to improve reliability by coordinating different projects to ensure dependent modules are updated simultaneously and work together so developers and ISVs aren't left hanging when new versions of Eclipse ship. Callisto in June wrapped up 23 projects run by 10 teams.

Another problem is quality. An Evans survey of 1,200 developers using 11 IDEs this summer put Eclipse last on quality with NetBeans. That may not be a great surprise given Eclipse was up against premier-level IDEs that have taken years to mature under the stable, project managed environment of a closed-source, enterprise Java vendor. It would explain, though, why Eclipse started Callisto and shows Eclipse's goal now should be refinement instead of extension.

And there's the lingering question over the degree of influence IBM retains. While a benign presence, IBM holds two seats on a board of 12, compared to everybody else's' one. More than 120 of Eclipse's 240 committers are from IBM, although numbers are being diluted as more companies join and new projects are created. And while plug-in providers are flocking to Eclipse, IBM remains the prime beneficiary through WebSphere, Rational and - now - Lotus.

"I'm never going to turn away committers - that would be silly. We are focused on growing the pie so there are more committers," Milinkovich counters.

Perhaps the first five years are just the end of the beginning for Eclipse. Now the hard work really begins.®