It's quite clear that I hit a raw nerve for David. The gist of my article is that I believe it is a flaw to continue to create a Java only object model when the technology exists to create an object model that all programming languages can access. JDO just further re-enforces the Java only, Java centric view of object model persistence started by the use of EJBs. JDO does not persist an application's object model in a database--it only persists the attributes of the object model. That's the rub. What about the behavior?
It's obvious that David has little knowledge about the SQL:1999 standard, and that Oracle simply has done the best job to date of supporting it. It's not proprietary object-relational mapping--it's object-relational technology--where both attributes and behavior are persisted as objects.
Even more obvious is that this article is really just a personal attack on the messenger. Ooowww, I had a typo in my class file in the article. If David went to the book store and actually picked up my book but failed to notice that Part IV of the book, three chapters, was all about the use SQL:1999 object-relational technology, including methods, with JDBC, then I guess he was just looking for a way to make a derogatory comment about my credibility. What can I say? David, get glasses and be a bit more thorough. Take your FUD finger and point it at yourself.
As Dion Almaer suggested in the feedback on my article, there is probably a place for all three technologies, an "object base" using object-relational technology, accessed by JDO for a data-entry program, or accessed by JDBC for a reporting program. However, the current implementation of JDO does not support OR methods, so it like JDBC will need to evolve through the Java Community Process in order to for us to achieve nirvana.