On the latest issue of eWeek , Mike Milinkovic, director of Eclipse Foundation, briefly compared NetBeans with Eclipse . Among other things, he said:


I'm always a bit surprised when I read "something is CLEARLY superior" in a scientific/technologic context, with no arguments about such a strong assertion :-O. Facts are expected here - and, furthermore, I don't expect that any major comparison between two excellent products such as NetBeans and Eclipse can ever be so clear to achieve unanimous approval, because there's always a subjective component; every heads see things in a different perspective. But even for making a subjective opinion one needs some facts.

For instance, my personal experience is exactly the opposite. To make a long story short, I've started focusing deeper on desktop technologies about four years ago, mainly by working on a project that initially started for fun and experimenting (my long tine background was more on the server side). At the time I was reather frustrated by the lack of advanced components in Swing and the lack of a framework - ending up in writing a messy framework of mine that was eating up all my spare time. Since at the time I was a happy Eclipse IDE user, I started studying Eclipse RCP since I felt it could help me. After a few weeks of experimenting I gave up, for a basic objective reason (I was frustrated at the thought of studying a completely new API - SWT - other than Swing) and some subjective reasons (I didn't like the Eclipse RCP APIs at all).

After almost giving up with my project, I gave it a last chance in 2006 when NetBeans 5 came out. Not only NetBeans 5 (today 6) gave me the capability of working with Swing, thus maximizing my learning ROI, but I found a rather good set of RCP APIs upon which I completely redesigned my application. I was able to use SwingLabs and the TimingFramework with no hassles, as well as customize the look and feel to taylor my needs (including the use of Quaqua to fit at the best the integration with Mac OS X). The point is that my application is about photography and targetted at end users, so a cool graphic design is as important as the software design (even heard of Romain Guy ?), especially if Mac OS X is one of the targets. If you look at the screenshots of some NetBeans RCP applications, you can see that mine stands out for its specific look and feel (that - BTW - it's far from being complete). And there's a pretty good choice in this scenario, as other third-party contributors provided configurable Swing Look and Feels with specific integration for NetBeans, such as Kirill Grouchnikov 's Substance . By contrast, all the Eclipse RCP applications I'm aware of share the same Eclipse look and feel - which is quite boring in some contexts (this of course doesn't detract a bit from their engineering quality).

So, two years later, my subjective adversion with respect to SWT has turned out to be backed by a more objective rationale: to me the capability of integrating Swing and third parties' Swing stuff is fundamental. I see that the same thing seems to be an important point for commercial products in the same field, such as LightZ one by LightCrafts which is written with Swing (and AFAIU - didn't try it - the latest version is based on NetBeans RCP as well). That's why I think NetBeans RCP is superior to Eclipse RCP.

Of course this is still my personal perspective, nevertheless is the kind of rationale I expect to support or confute a comparison. Others of course think the opposite - but may I know what are their rationales?