Tim'O Reilly, the eponymous publisher, kicked off EclipseCon 2005 in Burlinghame earlier this year with an excellent presentation titled "Open source business models and design patterns." As well as documenting various failures and successes in the computing world, one message that struck a chord was that to succeed in open source you must design for participation. Three days later, Lee Nackman, CTO of IBM Rational Software and one of the original thinkers behind the Eclipse project, demonstrated how this was one of the core principles built in from the ground up. It had led to the creation of the independent Eclipse Foundation that, among its other goals, had the aim of achieving serious participation from more vendors. This was demonstrated more than anything at the conference that, while last year was being talked about as a concept, occurred this year.

BEA, Borland, Scapa, and Sybase recently became strategic members of the Eclipse Foundation and are driving forward key projects. A quick look at the list of technology projects on www.eclipse.org/technology/ shows the level of participation and diversity of the efforts. The number of projects that were newly announced in the past few months is a nice litmus indicator of the health of Eclipse and its current growth phenomenon.

The booth floor of the conference had a large number of companies showcasing everything from testing and reporting products through code analyzers and high-level design tools. In 2004 many exhibitors were touting Eclipse integration while showing their existing product line elegantly shoe-horned into the workbench. This seems to be less prevalent where everyone from mom and pop to the big guns of software were demonstrating products that had a very high degree of fit and finish with respect to the platform and its integration API.

In 2004 there was a buzz around the ideas behind the Rich Client Platform (RCP); this year these were being demonstrated as a very real and working technology. The RCP's idea is that folks building tools for Eclipse should not be the only ones to benefit from its plug-in architecture, and developers of end-user applications in any domain can enjoy its concepts such as window management, UI frameworks, and having an update manager. RCP is an odd framework because on first inspection it doesn't really add anything to Eclipse; instead it allows for all vestiges of IDE-like features to be removed so the workbench becomes stripped down to its bare bones. What is impressive with RCP is that this idea has not only worked technically but has become an execution reality with projects such as NASA's Mars rover mission, which is planning on using it. The RCP team has also extended the plug-in development environment (PDE) to include specific wizards to help build RCP while the Visual Editor's SWT GUI builder supports creating RCP views. RCP often reminds me of the Internet, where a relatively simple idea is ready at the right time, free to use, and can have an explosive effect way beyond its initial conception on software development.

One of the questions at the end of EclipseCon 2004 was why the Web Tools Project (WTP) seemed to have stalled. Since then a lot has occurred with WTP and this year's conference had a number of technical sessions and a BoF on WTP showing its progress and the health of the project.

The Eclipse Test and Performance Tools Platform Project (TPTP) is another success story, with the goal of providing a common API and data model for monitoring, testing, tracing, and profiling tools. What is nice about TPTP is the way that its contributors have managed to create a very functional base yet they're the very same companies that provide value-add in their commercial offerings by extending TPTP in their own product lines that benefit from the core frameworks and data interchange model.

EcliseCon 2005 also saw interest in the Business Intelligence and Reporting Tools (BIRT), which has the goal of providing a set of reporting features and scripting to allow J2EE Web apps to produce Web- and PDF-based reports.

What I find refreshing about BIRT is that it was proposed by Actuate and TPTP is led by SCAPA, underlining the fact that the Eclipse Foundation enjoys and benefits from the depth and breadth of knowledge of the large number of companies that form its base.

I came away from this year's conference with the feeling that Eclipse, by its critical mass of participants, is being propelled along new and interesting paths that continue to stretch and test the framework along its underlying principle of extensibility around a plug-in architecture. One of the slides that stuck in my mind was by Carl Zetie, an analyst with Forrester's Application Development & Infrastructure Research Group. While doing market research on IDE usage, they had asked companies what their adoption of Eclipse was and received two different answers. The reported Eclipse usage from IT managers was lower than the programmers. The grass roots programmers who perhaps didn't like their department's set of tools had adopted Eclipse to become more productive, in a way that reminded me very much of the early growth of Microsoft Windows where individuals installed it and became productive while their IT departments were busy rolling out late and inflexible corporate systems. The fact that Jason Weber from Microsoft gave a talk at EclipseCon encouraging tools providers to develop for both Visual Studio and Eclipse simultaneously adds its own ironic chord.