C'mon, does anyone really believe stuffing compile/linking directives into comments is a good idea?
Thanks for the interesting article. Gave me alot of insights. Having worked extensively with one of the major J2EE vendor's product for the last five years I observe the biggest challenge to J2EE is the products in the space; thus, I cannot argue against someone using .NET other than to ask them to take a look at Open Source Java or even LAMP. The major J2EE products are productivity deadends. When I looked at JBoss it looked good so maybe it is better than Web@#$@#.
Stupid ideas like stuffing directives into comments are just the tip of the iceberg. If you want app-gen code, comment-annotation is not the way to manage it.
BTW, the crux of the article is sort of assinine. Spring is cranking. EJB3.0 is a pipe dream for most and lots more proprietary/risky for the bleeding edge products that already support it. Design by committee seems to be less optimal than design by teamwork.
Lastly, the POJO acronym is not so much an advancement as an illustration the original J2EE design was uninspired. Since writing the first EJBHome even though reflection was part of Java, all of us knew that. Talk about re-inventing IDL!