M. Holub on the JCP




The very respected Allen Holub has the following “Do We Really Need the JCP?” article on SD Times. He does have some points on the JCP process showing it’s not perfect but there’s quite a few things he misses IMO.




First, what other standards organization :


– brought more than 200 JSRs to the market?


– provides a specification, a test compatibility kit and a Reference Implementation for each JSR?




EJB today may be an unperfect technology (although this is the same old EJB = CMPs inaccurate “shortcut”), but no one is forcing you to use it. Actually the entry-level (i.e. affordable) products from IBM and BEA wouldn’t even let you use EJBs. J2EE never pretended you could develop without having an Architect in charge of the structure and technical choices. I think J2EE even made this job more critical. Note also that EJBs came out way before the JCP existed… Granted the JCP didn’t make it much better (until EJB 3.0).




I’m not sure what the JSF criticism has to do with all this. Maybe simply due to the fact JSF took a while to get there… It reminds me of the ongoing discussion I have with Anonymous Coward (although it’s most likely not M. Hollub ;-)




Let me also bring up one pretty obvious counter-example of the JCP’s achievements : Tiger (aka JDK 5.0).

M. Holub should know as he does consulting on the subject that this is a major achievement given the language evolution and the many other enhancements. It all started from day one per the JCP rules.




As for the Hibernate comment “I could support welcoming Hibernate into the JCP, but creating yet another competing technology is divisive and (I hope) an effort doomed to failure.”, M. Holub is missing the difference between a specification (EJB 3.0) and an implementation (Hibernate in this case). Maybe it’s also saying something about the way this effort should be advertised.




I think the Java community has a great place to incubate projects, it’s called java.net. This is a place where things like JDNC and JDIC (JavaDesktop projects) mature before they are eventually taken to the JCP.




So while M. Hollub has some points, I think he simply expects too much from the JCP.

Author: alexismp

Google Developer Relations in Paris.