Just read an article about the current complex state of affairs in the J2EE world. One of the comments to it seems to resonate among a lot of Java developers:

“J2EE is a great solution for a problem no one actually has. I can’t wait for something (anything) else to arrive that makes programming the simple, fun affair it was in the good ‘ol days.” —David Zimmel

In fact, when a friend of mine read the article, he said:

For my part, I use Python whenever I can: the platform is rich and complete and fast, and so far free of delusions of grandeur. And best of all, it’s fun.

I’ve heard his sentiments from lots of ex-Java people, and it got me thinking about the state of affairs, and whether my current view of software engineering is … well, fun.

I’ll admit that Java used to be fun when I started out, but eventually, the fun got lost for me for quite some time. Some of this was due to the fact that Java took off all over the place, and it became bewildering to keep up with all of the many huge Java projects going on. Other reasons were due to some poor decisions on Sun. And other reasons were due to a stagnation of the core of Java. Of course, a lot of my lack of fun in Java was due to my work conditions at the time, but I’m getting off-subject.

I’ve had a slight re-epiphany of sorts surrounding my re-discovery CBD. The concept works well from my perspective on extreme programming … the entire concept of unit testing small code sections work well if the structure is based on a component model where each component only interacts with other components through interfaces.

But somehow we have to manage the complexity associated with server-side Java. I concur that EJBs are an overly complicated mess and don’t get me started on the proliferation of “frameworks.” I think I’m just going to have to hark back to a simpler time and see if I can’t come up with a simpler approach.