Mike Milinkovich, executive director of the Eclipse Foundation, has been kind enough to answer some questions for Java Developer's Journal. Rather than rattle off the usual ones about the name, about why Swing wasn't used, or how much influence IBM still has, Mike has fielded questions on some more current and topical subjects, as well as given us his insights onto the future. Thanks for taking the time to talk to us Mike.
JDJ: The Eclipse Foundation recently joined the Java Community Process. Can you tell us how this is going and what you expect to get out of this, as well as give to the JCP?
Mike Milinkovich: Yes, we recently joined the JCP, as we also joined the OSGi Alliance and OMG. The reason for joining these organizations is that the Eclipse community relies heavily on the standards that are produced by these standards organization, so we wanted to show our support.
As for the JCP specifically, we are planning to contribute in a couple of different areas, the most immediate example being the use of the Eclipse Equinox code as the reference implementation for JSR291.
JDJ: At this year's EclipseCon I felt that the amount of interest in RCP had surpassed the amount of interest in the actual IDE. Do you think this is the case, and if so does this change the dynamics of Eclipse's strategy and direction to become more of a general-purpose application platform and less of a development environment?
Milinkovich: Yes, I agree the Eclipse community and the industry as a whole has moved toward viewing Eclipse as an application platform. We are seeing a lot of interest and adoption of RCP and also projects such as Equinox, RAP, and Higgins. However, this is not new as we have had a conscious strategy to move Eclipse beyond just being a Java IDE for several years now. I believe what we are seeing is quite simply that a number of the newer projects within the Eclipse community are becoming more mature and are enjoying greater interest and adoption as a result. The vision for Eclipse has always been about being a complete platform for software development and I think we are well on the way.
However, I do continue to see a lot of interest in Eclipse as a tools platform. We have new projects for providing IDEs for dynamic languages such as Ruby and PHP. The AJAX Toolkit Framework (ATF) is attracting a lot of interest as a tooling platform for AJAX developers. CDT, our C/C++ IDE, has great momentum as being the C/C++ IDE for embedded and Linux development. In my opinion, Mylar is one of the most innovative new developer technologies to come about in a long time.
Therefore, I don't really see a large change in strategy or direction; I see it more as a journey and evolution. This is what makes Eclipse such a vibrant and interesting community.
JDJ: JSR 291 ratified the OSGi Java module mechanism used by Eclipse to become part of the Java language specification. Can you see the same occurring with SWT?
Milinkovich: I haven't seen any interest from the community in putting SWT into the JCP process. It is not that I don't think it would be an interesting idea but someone would have to step up to spend the considerable amount of time required to take it through the process.
JDJ: How do you manage the relationship between the Eclipse board members, some of whom are fierce competitors in the commercial marketplace, yet need to collaborate for the good of the platform?
Milinkovich: Interestingly, this has been mostly a non-issue to date. I've been really very happy with how collegial and effective the Board of the Eclipse Foundation has been.
That said, I think you have already identified the answer; there is a common need for a strong, stable platform for building all sorts of different software. Eclipse is providing this platform and it really becomes a unifying force at the Board and through-out the community. The other thing is that the Eclipse governance model is proving to be very good at managing the different interests that participate in Eclipse. All the organizations have an equal say at the Board level and the principles of meritocracy and transparency help resolve a lot of the issues within the projects.
JDJ: What excites you most about what is going on with Eclipse at the moment?
Milinkovich: Europa, our next release, is going to be pretty exciting. We have over 22 projects lined up for the annual release train, so a lot of new stuff is getting ready for release. Our annual release trains are very important for the entire Eclipse community. First of all, they are a real testament to the committer and project community's ability to deliver on a predictable schedule. Second, they are hugely important to our adopter community as they use the projects to deliver their commercial products or open source projects.
I think the growing involvement of Eclipse in the world of Equinox-based OSGi runtimes - such as the EclipseLink persistence project recently proposed by Oracle - is cool as well. Over the next 12-18 months we are going to see a lot of new stuff being built with Equinox.
JDJ: What worries you most about what is going on with Eclipse at the moment?
Milinkovich: I'm not sure that "worried" is the right word, but I would like to see vertical market frameworks (e.g., banking frameworks, health care frameworks) being developed as Eclipse open source projects. This will require enterprises to become more involved with contributing to open source projects and it is something that I think will take time. We have a great start with the Eclipse Open Healthcare Framework (OHF) but we need to do more to encourage large enterprises to begin collaborating in open source projects.
JDJ: NetBeans seems to be gaining a lot of traction at the moment, especially with some of the emerging and BRIC markets. What is Eclipse doing in this space to keep up?
Milinkovich: This is a funny question since we are finding the BRIC countries to be our highest growth areas. Our downloads from China are just exploding; Evans Data recently reported our usage grew 30% in India and 20% in Brazil. It is nice to hear that NetBeans is doing well but I certainly don't feel like we need to keep up. In fact, I believe we are leading in the BRIC countries.
JDJ: There is talk that Eclipse 4.0 will be a major rethink of the platform and how it is put together. Can you share with us what's going on here?
Milinkovich: To be honest, everything about Eclipse 4.0 is conjecture at this point. EclipseCon was the first time the idea had even been discussed by the community and not only are there no concrete plans yet for 4.0, there isn't even yet a firm commitment that it is going to happen. Of course, I think it will, but it is definitely too early for me to speculate what may or may not be in it.
What is going on at this point is the beginning of a process - a community process to decide whether doing a major new release - quite possibly with some API breakages - is the right thing to do for our broad ecosystem.
JDJ: If you could do Eclipse all over again, what would you do differently and why?
Milinkovich: First, please remember, I did not "do" Eclipse. A lot of people spent a considerable amount of time and energy establishing Eclipse long before I appeared on the scene in 2004. I have just been lucky enough to be involved in the implementation of the vision they set out.
Things are working remarkably well, so I am not sure I would change anything. It is my belief Eclipse has become a place where many organizations can collaborate and innovate on interesting new technology. I truly believe Eclipse has the right technology, architecture, and governance to make it an important platform for future software innovation for many years to come.
Bill Roth has had a distinguished career in Java, formerly being the JEE spec lead at Sun Microsystems where he worked on some of the early EJB specifications. In a previous life, Bill was a contributing editor to Java Developer's Journal. He maintains a lively blog and is currently vice president of the Workshop Business Unit at BEA. Bill has kindly agreed to answer some questions on BEA, Java, and all things interesting to do with the two.
JDJ: Can you tell us a little about what's going on at BEA at the moment?
Bill Roth: There is a lot going on in Java right now at BEA. First off, we are in the final stages of delivering the next generation of BEA Workshop. This will allow the construction of blended applications, the mixing of commercial source and open source. This will allow developers to take advantage of the best of both worlds, in the way they want. I call it our "Burger King" strategy, where developers can "Have it Your Way."
We're also working on an exciting new vision on how all the roles in IT can work together. This effort, WorkSpace 360, is a vision for unifying the various participants across the SOA life cycle.
WorkSpace 360 is intended to break down the communication barriers that exist between the various participants in the SOA life cycle: business analysts, architects, developers, and IT ops. It provides the ability for the various participants to share information and assets among each other in a seamless, governed manner. This is enabled through a series of tools, communication capabilities, and views tailored to the individual stakeholders. At the core of WorkSpace 360 is a centralized metadata repository that serves as a central source of record for enabling the seamless flow of information across the different stages of the SOA life cycle.
JDJ: Recently there was some ruckus in the press about BEA failing to announce third-quarter 2006, which meant a possible delisting of their stock. What happened here, and is there any long-term damage?
Roth: We are among the over 200 publicly traded companies who are reviewing their options granting practices. BEA has been, since the beginning of the investigation, working collaboratively with the SEC and the stock exchange, and we continue to keep them in the loop every step of the way as we make progress. From everything we have seen, NASDAQ has worked constructively to avoid delisting when the companies involved are working in good faith to resolve their issues and get financial statements back on file as quickly as possible. We believe we fall into this category.
JDJ: For JRockit, do you see yourself as being able to compete commercially with Sun, whose JVM is freely available?
Roth: Of course we can. Not only is JRockit freely available, but has been shown to be 24-28% faster than the leading JVM. We'll compete with Sun's JVM any day of the week.
JDJ: The app server market is becoming largely commoditized with open source projects like JBoss, Tomcat, or Geronimo. How can WebLogic compete in this space and remain relevant?
Roth: Your question is based on an erroneous assumption. While certain segments of the application server market are indeed commoditizing, saying the entire market is "largely commoditized" is overreaching. If this were the case, our WebLogic business would be shrinking, and it most certainly is not.
JDJ: Is BEA aligned very strongly to Java as a server-side programming model, or are you embracing things like PHP?
Roth: I have blogged about this in the past. While we're committed to the J2EE programming model, it's clear that developers are looking at other technologies. We have a number of projects where we're working with next-generation dynamic languages like PHP and Ruby, so we can be ready as the enterprise developers begin adopting them to build their applications.
JDJ: What do you think of Web 2.0 and all the AJAX excitement? Is this something you're tooling for and adopting in WebLogic?
Roth: AJAX is an exciting new area of great technology, but it is also an area of great chaos as well. There are way too many AJAX frameworks and no standards that are clearly emerging. Also, no one has clearly articulated a declarative, standards-based XAML-like way of defining the UI for this technology. As such, our strategy is to work with a small number of vendors in the short term to deliver value to our developers, and then keep an eye on the standardization process.
JDJ: What excites you most about what is going on with BEA at the moment?
Roth: What excites me the most about working at BEA is the pace of innovation and our plans for the next five years. When you couple the success of AquaLogic along with our vision for Workspace 360 and SOA 360, I am even more convinced than before that BEA is on the right track.
JDJ: What excites you most about what is going on with Java at the moment?
Roth: Same answer, really. The pace of innovation in the area of open source software and frameworks is really exciting. We're also seeing a shift in user requests to JSF from Struts, but it also appears that Struts 2.0 is picking up steam as well. What's truly unique is that the bulk of the innovation these days appears to be happening outside of the Java Community Process, and I view this as a good thing. The Java Community is growing organically in ways that are impossible to predict.