Marc Logemann bemoans what he percieves as a lack of progress in JSR-198 (the standard API for writing Java IDE extensions). Rumors of JSR-198s demise have been greatly exaggerated ;)
One effect of the way the Java Community Process (JCP) works is that, unless you're in the expert group for a particular JSR, you won't see a lot of information about the progress of the JSR until the point it goes into public review. I don't think this lack of transparency is a good thing. It leads to the kind of confusion that Marc has right now: you have no idea whether JSRs are stalled completely or under active development.
I don't speak officially for the expert group in any way, and I want to be clear that I'm speaking personal opinion in this entry and not reflecting the opinions of my employer in any way. I can say from direct experience that the JSR-198 EG mailing list is pretty active. The spec is edging closer to public review. It's currently circulating inside Oracle for internal review. Following this, it will be submitted to the expert group for their review.
Implementation of JSR-198 support is also proceeding both for JDeveloper and NetBeans. In the case of JDeveloper, we hope to have full support for JSR-198 in our upcoming 10.1.3 production release. We're investing so heavily in the technology that we've made it our default extension mechanism. When you build an extension for JDeveloper 10.1.3, you're building an extension that can potentially be used in any JSR-198 compatible IDE.
In light of that, it was interesting to read in the comments on Marc's entry that Eugene Belyaev, President and CTO of JetBrains, has no faith in JSR-198's future at all:
Yes, we joined the EG upon foundation, however, later realized that this initiative will never work as planned.
Eugene is not specific about why JetBrains rejected JSR-198. I've only recently been involved in the 198 EG, so I'd be interested to hear whether these were technical issues with JSR-198 itself or issues with the extensibility of IntelliJ. In any event, it will be telling to see whether their stance on the issue changes after we have working implementations in NetBeans and JDeveloper.
I can honestly say that I think 198 is not by any stretch of the imagination a perfect solution to the extension development problem, so I can understand that companies like JetBrains have concerns about it. That said, the thing that fundamentally surprised me about this is that they seem to have rejected the idea of a common extension API out of hand rather than collaborating with the EG toward a solution that could work. I can't really think of a strong motive for this. It's not like JetBrains have an overwhelming share of the non-Eclipse IDE market such that they can afford to go their own way.
Marc ultimately makes a good point:
I find this JSR ... important because plugin architecture is the only plus of eclipse and by unifying it in IDEs like IDEA, Netbeans or JDeveloper ..., we would get ... some more competition and inovation in the IDE area.
Eclipse is extremely strong because of its plugin architecture. One argument I've heard expressed against JSR-198 goes along the lines of, "why do we need a common extension API? Why not just make Eclipse the standard IDE?". Java IDEs have flourished over the last few years precisely because of the competition between IDE vendors. I may be biased, but an Eclipse-only world is not the choiceless future I want. I'm sure I'm not alone...
