Late last week, Oracle announced it would no longer be developing the Ruby on Rails (RoR) plugin for the NetBeans IDE. To be explicit, Ruby and RoR will no longer be included in the upcoming v. 7 release of NetBeans. There are two ways to look at this: It's overdue, or it's another symptom of Oracle's cavalier view of open-source projects. I tend to lean more towards the former.
I've closely followed the Java IDE market for years. Not only as a Jolt judge, but as a reviewer. During the last six years, I have written several long, comparative reviews of the major Java IDEs, including one just a few months ago. In these reviews, my consistent perception was that NetBeans was the best of the free OSS products. It was faster and more friendly than Eclipse and Oracle's JDeveloper, and it had good support for a broad range of languages. I also liked Sun's commitment to the product. After badly neglecting the Java IDE space, five years ago, Sun decided to get back into the game and convert NetBeans from an also-ran into a contender. It rewrote the editing module and then added a whole array of features. One of those was support for Ruby and RoR. If I recall correctly, it was the first of the Java IDEs to support Ruby. And it did so quite well. (At the time, I was surprised that Sun chose Ruby, rather than Groovy. The product-marketing manager at that time made it clear the decision was based solely on popularity — RoR was really taking off then.)
Today, however, Ruby enjoys several excellent IDEs, some of which(say, JetBrain's RubyMine) are dedicated Ruby environments. While NetBeans still has very good RoR v. 2.x support, it seems increasingly unlikely that someone would opt to use a Java IDE to do Ruby development. (This argument is undermined somewhat by JRuby, which is best served by precisely such a hybrid platform.) Moreover, the Ruby community inside NetBeans has been steadily dwindling. Prior to the announcement, only 15 threads were posted to the mailing list this month, of which roughly half did not receive an answer. One query was the deadly "Is this project still alive?" post.
Part of the problem is that one of the lead programmers behind the Ruby support is Erno Mononen, a Finn programmer (recall NetBeans' core team is Czech) who, after working at Sun and Oracle, has moved to a Finnish IT consultancy. None of this augured well for the IDE.
The breaking point came when NetBeans v. 6.9 could not run RoR 3.0 correctly. At this point, I believe, Oracle declined to tie up one or more developers to learn the plugin's codebase and make the fixes that would keep RoR on NetBeans moving forward. In light of the decreasing usage, this seems like a reasonable decision. Further contributing to the decision may be the fact that RoR is hardly a target market for Oracle, so there was no strategic goal under which such an expenditure could be justified.
The question that disturbs me more, however, is what other parts of NetBeans might be sheared off in the future? The question is hard to answer. Oracle's minimally forthcoming attitude about OSS plans makes it difficult to assess. However, the company did keep NetBeans going, even though it has its own IDE in JDeveloper and maintains a separate Eclipse package. And, indeed, Oracle is underwriting the upcoming v. 7.0 release of NetBeans. So, it has certainly not been resistant to the project's progress.
But there is a cost to the decision beyond loss of the Ruby aficionados. This is not the first time NetBeans has lost an important feature. Last year, support for UML modeling was removed. Now Ruby. The question of what else might be excised is important along one central vector: adoption by new users. Committing to a new IDE is a significant undertaking — from learning how the environment views projects and components, to muscle memory of new keystroke combinations. To induce developers to make this commitment to a new product, they must believe that features that attracted them will remain in the product, if not improved in subsequent releases. Without this, an IDE risks slipping progressively into irrelevance.
(Update: This story originally reported incorrectly that Erno Mononen was Czech.)