I just got through viewing an Eclipse Seminar hosted by Raghu Srinivasan which provided a live demonstration of what you can do with Eclipse 3.3, Web Tools Project (WTP) and Java Server Faces (JSF). The live demo did a good job of illustrating the productivity boost that the Eclipse platform is capable of providing for enterprise Java development tasks. Despite time constraints, the ease with which you can pull together anything from a simple, functionally sound web app to a complex enterprise architectures, was clearly evident. Eclipse continues to set itself apart as an open, integrated development platform by consistently strengthening a series of complementary toolsets, plug-ins and integrated offerings. We are only beginning to see the fruits of what has been a consistent, focused approach to building not only a product, but a community coupled with an open governance model (The Eclipse Foundation).
After starting out as an IBM contribution to the open source space it has grown into a full fledged ecosystem fit with contributors from across the globe. Plus it provides a great example of how open source can create added dimensions of value outside its imminent domain. Just look at Hibernate and Spring, both of which are great open source projects in their own right, and how they have benefited tremendously from the existence of a solidly architected Eclipse Java development environment. Having strong, open source lightweight J2EE frameworks and persistence mechanisms without an equally capable open platform to serve as a tie-in, doesn't prove as 'sticky.' Not to say they couldn't exist on their own, but it's very easy to identify the symbiotic relationship.
During the early stages, I always pictured IBM always keeping Eclipse under tight wraps as a free tease for what has become their Rational Application Development suite...obviously I was totally off the mark. Eclipse is steadily evolving into a monster of a value generator for more than just IBM, all the while sticking with an open source approach.
On the other hand, I would still like to see more done towards establishing a better web delivery strategy for the growing catalogue of Eclipse based open source software. Yes, you can currently download anything you need from eclipse.org but you have to do so in discrete steps which tends to cause confusion when it comes to dependencies. A slick Web 2.0-esque app which enables you basically customize install packages might really take off, especially if it did things like list required libraries and/or provide a clean view of available projects, components and tools. Perhaps the community could charge a nominal subscription fee to significant offset bandwidth/administrative costs, and further down the line it might be capable of doubling as an entry point into some sort of Eclipse Developers Network.
The objective of doing so should be to continue to find ways to sharpen not only the quality of the actual software but also the according user experience. In the meantime, it's going to be interesting to see how Eclipse plans to translate impressive growth + gathering momentum into their agenda for 2007.