I lost my temper again

Am I really the only one who thinks EJB 3.1 is a case of jumping from the mud into a puddle? I mean no harm by that. A puddle is better than mud, but I can hardly call it a leap forward. EJB's problem is that it can never shake the stigma of “we know what's best for you, and if you disagree, you're doing something wrong.” I lost my temper again while reading EJB 3.1 — EJB New and Improved! I'll be delighted the day EJB starts innovating instead of trying to catch a technology train that left long ago. Alas.

What bothers me most is the failure to understand the basic requirements of a DI, or if you prefer, IoC container. It is a good illustration of how the people around EJB still fail to grasp what it means for a component framework to stay out of your way and avoid intruding on your code. Rod Johnson explained it to them nearly five years ago in his book J2EE Design and Development. What's more, he has built a remarkable career in the meantime by putting that idea into practice.

EJB's only hope may be to catch up with what is common elsewhere and then offer something extra. That might be OSGi, for example, or some form of hot deployment. Otherwise, it will always be the standardized but second-rate cousin…