"We continue to struggle a bit with what developers think “Eclipse” means. They have heard of it, but they believe that we are entirely focused on Java tools when in fact we are doing so much more," says Mike Milinkovich, Executive Director of the Eclipse Foundation, in this exclusive Q&A with Jeremy Geelan. "Our goals at Eclipse are to create an industrial-strength open source development platform that spans extensible tools, frameworks and runtimes," adds Milinkovich - pictured here during a previous Webcast on SYS-CON.TV from our Times Square studio.

Eclipse Developer's Journal: May 20th marked your 4th anniversary as the Executive Director of the Eclipse Foundation. What have been the biggest changes in the Eclipse ecosystem in that time?

Mike Milinkovich: I believe the biggest change is the breadth of the projects that are happening at Eclipse today, and the ecosystem those projects are enabling. Four years ago, Eclipse was clearly focused on tooling and the Rich Client Platform was just in the process of being released. Now we have a large number of really exciting runtime projects happening at Eclipse such as Equinox, EclipseLink, Swordfish, Rich Ajax Platform and Higgins. When middleware products like WebLogic and Websphere are both running on top of Equinox’s OSGi container, it is hard to argue that Eclipse is only going tools!

The other major change is that within the tools space, Eclipse has really broadened its reach from its original focus on Java. Today, the Eclipse ecosystem has a large footprint in areas such as modeling, embedded, mobile, RIA and Ajax, SOA and dynamic languages. If you are a software developer, you can almost certainly find Eclipse-based tools to help you.

EDJ: So do you no longer characterize Eclipse as "open source’s best kept secret’?

Milinkovich: We now regularly see Eclipse included in the list of prominent open source communities listed in press and analyst articles. You know what I mean: articles that say things like “…open source communities such as Linux, Apache, Mozilla and Eclipse.” So I would agree that Eclipse is no longer open source’s best kept secret.

That said, we do continue to struggle a bit with what developers think “Eclipse” means. They have heard of it, but they believe that we are entirely focused on Java tools when in fact we are doing so much more.

EDJ: I know you’re a fan of Strategy as Ecology. Using that HBR article’s typology, can you unpack the Eclipse status quo in terms of: Productivity of the Ecosystem, Robustness, and Niche Creation?

Milinkovich: So for those unfamiliar with the article, it basically describes those three measures as a way to determine the health of an ecosystem. Unsurprisingly, I think that Eclipse is doing pretty well.

“Productivity of the Ecosystem” refers to how much value is being created. And since this is a business journal article, it is referring to business value. Right now, I would say Eclipse is doing quite well on that dimension. There are literally hundreds of companies building thousand of products based on Eclipse open source technologies. In terms of market share, the combination of Eclipse and Eclipse-based developer tools is just about neck-and-neck with Visual Studio. So obviously, there is a great deal of value being created in the Eclipse ecosystem.

“Robustness” means how durable the ecosystem is and how well it adapts to external events, particularly those events which have a negative impact. On that dimension, I would say that Eclipse is so far doing OK, but not great. We have several examples of where the backers of an existing project have had – for whatever reason – cause to reduce their investment. In several cases (CDT in particular), the community has responded well and the project is now in an even stronger position. In other cases, the project has died (anyone remember Stellation?)

“Niche Creation” refers to the ability of the ecosystem to grow and expand into new areas of opportunity. This is an area where I believe Eclipse is truly outstanding. There are new Eclipse open source projects and Eclipse-based projects being started almost weekly, and many of those are exploiting opportunities created by new technologies such as RIA, Ajax, modeling and the like. The economics of using Eclipse as a free platform for developing new tools, frameworks and platform is just too compelling for both hobbyists and businesses alike to ignore.

EDJ: What do you usually say to those who assume that the Eclipse Foundation is about industry-wide interoperability rather than creating a development platform?

Milinkovich: Our goals at Eclipse are to create an industrial-strength open source development platform that spans extensible tools, frameworks and runtimes. And we want that development platform to be built by many individuals and corporations working together. It is a pretty ambitious goal, but so far it seems to be working.

EDJ: Isn’t “development platform” itself a bit vague as a term?

Milinkovich: Happily, yes. But how we have interpreted it so far is that our projects at Eclipse share a common architecture: the OSGi-based plug-in model All projects at Eclipse share this architectural invariant, and as a result each new project at Eclipse immediately adds value to, and profits from the others.

EDJ:  Would you care to expound on your notion of “the stackless stack”?

Milinkovich: That term was actually first coined by Michael Coté at Redmonk. But basically the idea is that the next generation of runtimes is going to be component-based, but in a way we have never seen before. All of us who have been in the industry for a long time have become somewhat jaded by the use of “components” and “component models” as some sort of panacea for developers. I certainly have. But there is something new about how Equinox works: the component model does not end at the container. In fact, almost all of the container is and its runtime services are implemented in terms of that same component model, right down to the very small bootstrapping runtime at the bottom of the stack.

In previous generations of component models, the container was a monolithic implementation that had to always be deployed in its entirety. With Equinox, developers will have control over which runtimes services they want to deploy with their application. And they can use those runtime services exactly like any other component – there is nothing special about “runtime” services versus “application” services.

The goal is to provide a much simplified and more consistent component model, along with the ability to provide highly tailored runtime solutions to match the application requirements.

EDJ: Is there any one particular project that comes to mind that especially characterizes what you believe to be best and most important/useful about Eclipse?

Milinkovich: I constantly use our C/C++ development tools (CDT) project as an example of a well run open source project and a great commercial ecosystem. Doug Schaefer (Windriver by way of QNX and IBM) has been leading the project for a number of years and has done a great job of attracting and keeping a diverse set of committers. Many of those committers work for competitors of one another.

On the commercial ecosystem side, CDT has become amazingly popular. It provides the basis for the tooling for almost every single real-time operating system on the planet, along with a significant percentage of the mobile market. There are well over a hundred products shipping that are based on CDT.

As an example of an industry collaboration that has supported a transparent and diverse open source project and spurred many commercial products, CDT is hard to beat.

EDJ: When the Eclipse Foundation turns 5 next January, what would be the best birthday present you could imagine?

Milinkovich: A developer survey that shows that Equinox’s OSGi-based runtimes are on the evaluation list for the future technology platforms of many ISVs and enterprises.

EDJ: Does your fondness for science fiction help or hinder you, when it comes to being active on the real-world software stage in 2008 and beyond?

Milinkovich: Lots of engineers and hackers are sci-fi fans. Imagining the future and wanting to help create it is almost the definition of our profession.

EDJ: How about ice hockey – are there an analogies there that you can usefully share as a way of shedding light on the software development community?

Milinkovich: Well, the first thing to note is that no one in Canada calls it “ice” hockey. There is only one type of hockey in this country! ;-)

There is almost nothing more satisfying in life than finishing your (body) check. Especially against the other team’s prima donna. Just watch what Detroit’s defensive squad has done so far in the Stanley Cup playoffs against Sidney Crosby and Evgeni Malkin of the Pittsburg Penguins.

In other words, it’s not the marketing flash that wins in the end. Delivering solid technology with a full team effort does. You do need your star players, but everyone on the team has to contribute to win.