I have to say, I believe in JPA. Being a former BEA consultant, I used entity beans at a very early stage of the spec and was one of the first to do a prototype of entity beans 2.0 for a customer with a beta version of weblogic (6.0 if I still remember back in 2001). I thought that wasn‘t very good (to be polite). We had customers doing so many funny things with entity beans that it couldn‘t be right. The spec was too difficult and customers didn‘t have a clue to use it the right way. And then Hibernate arrived, and now JPA.
I believe the JPA spec will be the one in terms of ORM because it‘s inspired by excellent frameworks such as Hibernate or TopLink, uses annotations which gets rid off XML config files and it looks like everybody around the JCP table agreed on it (not like JDO, R.I.P). Of course there are still things to be improved, but for a 1.0 it‘s not bad at all.
I‘ve written a first article about JPA (second coming soon) for DevX. It explains the basic mapping annotations, one-to-one relationship, Entity Manager and how to use JPQL. The second is more advanced and deals with inheritance, one-to-many, many-to-many relationships and JPQL. I‘ll let you know when it‘s out.
I would like to thank Thierry Balla, David Dewalle, Matthieu Riou and Thomas Verin for reading the articles before they got published and giving me some feedback.