The following is a transcript of the panel that took place during the Java Application Server Day in Genova.

Panelists:

• Paolo Ramasso (Oracle)
• Alessio Soldano (RedHat)
• Alef Arendsen (SpringSource)
• Alexis Moussine-Pouchkine (Sun Microsystems)

Moderator: Paolo Predonzani

Question: This year is going to be the tenth year since the introduction of Java Enterprise Edition. In your opinion, in all these years what has been the most significant contribution of your company to Java enterprise software?

Paolo Ramasso: Oracle has been a player in Java EE for a long time. In ten years we had a number of significant acquisitions, including Orion and BEA, and internal developments of the platform. I think the focus of Oracle has been to provide a complete solution for the enterprise that is scalable and ready to address the most challenging systems. This was done through the strategic acquisitions that I mentioned before, but also through other developments and acquisitions that grew little by little to build the most performant open standard-based solution on the market right now.

Alessio Soldano: Contributing to Java enterprise software means two things. On the one side, it means providing something that the enterprise can actually use to perform its business. On the other side, it means collaborating to the definition of the standards on which the enterprise platform is based. JBoss, even before the acquisition by RedHat, has been focused on both these aspects. You saw from my presentation that even the first versions of JBoss provided support to and compliance with the Java EE standard. Also currently JBoss/RedHat is actively contributing to several expert groups and JCP's. Collaborating in the definition of standards means shaping the future of the Java enterprise platform.

Alef Arendsen: Not in the specific area of Java Enterprise Edition but more generally in the area of Java enterprise software, there are two projects that have been most groundbreaking: Hibernate and Spring. Both did innovation outside the standards bodies and focused on standardization only later. Our main contribution has been Spring itself, which took away the complexity and bloat out of Java EE, and from which Java EE later took inspiration.

Alexis Moussine-Pouchkine: Of course Sun has been driving the standardization of the Java platform, working with the JCP, and providing many of the reference implementations. The standard itself has a lot of value. You can realize that when your historic vendor changes something in what it's doing - like raising its maintenance costs - and you need an escape door: this is where having a standard is absolutely key. We took inspiration from several other technologies. Java EE 5 owes a lot to Spring but also to other technologies. At a product level, we have an application server (GlassFish) that is a real product, not just a reference implementation.

Question: You mentioned the Java Community Process and the importance of having standards. But also innovation is important. Can you explain how you approach these two aspects and how you find a compromise between them?

Paolo Ramasso: I can answer with a specific example, Oracle In Memory Distributed Data Grid, which is an implementation and an extension of JCache. We did a lot without waiting for the standard to be completed. We introduced innovative features that are not standard but built on top of a standard baseline. Something similar also happened a while ago with object-relational mapping. There were Oracle TopLink, Hibernate, which were innovative and proprietary. Then things moved in the direction of standardization.

Alessio Soldano: I think most innovative things start outside of standards, but later have to undergo a process of standardization. Innovative technologies are good but it is also important to have technologies that interoperate with each other. I personally work in the web services area and you know that web services are all about interoperability. Without standards there would be no interoperability, but also without innovation and new features there would be no progress. That's the case for instance of JSR-299 [Web Beans], which started proprietary but is moving toward standardization.

Alef Arendsen: On one hand, I believe innovation doesn't start inside standards bodies. "Thinking outside of the box" is not possible if you are with many other vendors. On the other hand, if we think of innovation as improvement, then I think it is valuable to standardize certain basic features. In that sense there can be innovation within a standardization body. From our perspective, standardization comes from compromise, and that's pretty tricky to manage. In the JCP specifically, many compromises have been made. Things in the JCP are improving however. Also I believe it is an interesting thing to have more discussion in the community, before standardization, as a means to prevent compromise.

Alexis Moussine-Pouchkine: I think standards don't have to start necessarily from a blank page. The first version of EJB started from a blank page, in the sense that IBM brought in their expertise with business objects and transactional processing, and shaped the standard. But that doesn't have to be the case all the time. Discussion and the community can be part of the standardization process. In fact the JCP is a space where the discussion can happen. However what we see is that, even if a space for discussion is present, the community sometimes doesn't participate, at least at the beginning. In these cases a standard can be born without much discussion and only later the community joins in and starts suggesting improvements. If you care about Java EE you should care about the JCP and participate.

Read more Application Server day reports or have a look at the slides.