In this interview, Tim Boudreau said that an important consideration in evolving NetBeans is longevity.

In the past few years, the leading IDEs, including NetBeans, have grown bigger and more complex. That should not be surprising, given that "I" in IDE stands for "integrated." However, if that trend continues, at some point entropy will ensue, making IDEs too cumbersome, and less usable for common tasks.

Where do you see IDEs heading in the coming years? Do you see simpler tools that perform a handful of tasks well, or do you see room for more integration inside comprehensive IDEs?

http://www.artima.com/lejava/articles/netbeans_platform.html

I'm a big fan of IDEs, but I find the trend of making them into general purpose platforms (RCPs) disturbing. I think the path to an optimal IDE and the path to an optimal platform are not necessarily the same.

I also think there is a temptation to inappropriately incorporate these platforms into one's own applications. The result will probably be a much large code base then would otherwise be required. Given the time to learn these platforms, using them won't always save time either.

I see this platform idea spill into more modest projects where the designer insists on creating a more general and time-consuming solution than is needed in the hope that some of it will be reused later, only to have the company change strategies or go out of business before any reuse can occur.

I'm not against all general purpose solutions, I just think that currently we are out of balance.

There have four leading java IDEs,Netbeans,IDEA,jbuilder(SOLD) and eclipse. But in fact there have many lightweight IDEs,such as jcreator.
I like lightweight IDEs because all of them are very fast.
I don't like eclipse,bacuse I have to search tons of plugin for it.
IDEA is suitable for me,but I have to pay $380 for it.
Then I have to use netbeans when I develop web application,