Alex Miller: I was happy to see that InfoQ took up the JSR 277 / JSR 291 stuff again and did a pretty good job weaving together the recent activity (including my own call to interest on the subject).

...

So I then have to ask whether JSR 277 is re-inventing the wheel (to some degree) only to split the market and hurt the overall community. This is a somewhat rhetorical question of course. I believe that everyone involved here is truly trying to build something great. But sometimes I think we need to step back and see the bigger long-term picture and ask some questions about our goals.

Somehow I feel I've heard this story before:

• Some feature was developed outside the JDK to fulfill a need. And then a similar but different feature is incorporated into the JDK. The first example of such a feature is the Collections API. It offered functionality similar to a commercial product whose name I can't even remember. A second example is the Java Logging API, which in theory should have eclipsed Log4j by now. But Log4j is still being used.
• A component system is introduced into the JDK with the hope that it will create a vibrant third party component market. The promise is always the same: you can go out and buy a commercial package of ready made components and just plug them in to your project and development time will be cut in half. On top of that, you can assemble your applications with a GUI and a hundred drag-and-drop monkeys. The JavaBeans specification is the first attempt at this. (But how many JavaBeans did you buy last year?) The EJB specification is the second example. It is a huge win—for IBM and BEA—for a few years (and we are all saved by JBoss and then Spring and Hibernate). But did anybody sell any ready made EJBs that can be just dropped into any application server and just worked? And did you remember to turn off the light when your stateful session bean passivates?
• And then there is the entity bean vs. session bean story. I've heard this story in one of the earlier NFJS symposiums from Bruce Tate. In the early days of the EJB specification creation, there are conflict proposals for how an EJB should be implemented. One proposal, by Oracle if I remember Bruce's story right, is centered around a relational database. Another propose, I forgot by which company, could be Sun or IBM, has a different proposal that does not involve a database. And the EJB 1.0 specification included both as a compromise. And neither was adequate.
• The Java Generics feature was debated heatedly back in the days just like the current debate, with erasure at the center of the debate. And the wrong side won. As a result, the JDK now contains a Generics feature that is less elegant than the similar feature in C#/.NET.
• The JDO people yelled and screamed at the EJB people when JPA (EJB 3) was being proposed. To a lot of JDO user's mind, EJB 3 was reinventing a lesser round wheel. But EJB 3 prevailed.

I have no doubt that both sides of the current debate are trying to make Java based systems more modular. But I don't know how the current debate will play itself out. I just wish those involved, especially the big vendors (Sun and IBM, etc.) would put the interest of the community in front of the interest of their company and produce something that's unifying and useful.

Responses[7]

Posted by Weiqi Gao on September 14, 2007 8:41:51 AM CDT #