Jorgen Thelin (Microsoft) and Tim Bray (Sun) are talking about how standards are terrific when applied to proven industry practice but high-risk in the domain of theory and science.
I definitely agree. If we take a look at the JCP what are the stars, and what are the dullards?
java.util.concurrent
This bad boy came out of the great work and experience of Doug Lea and his concurrent package. It evolved, it was proven in the real world (which Jorgen likes), and was then just fine-tuned in the standards process. In fact it gave Doug a chance to take his experience, but also "start again" and clean house.
Java 1.4 Logging
The de facto standard at the time was Log4J. However the JCP seemed to ignore that and make their own thing. Log4J even had to change to look more like the jdk 1.4 logging API!
JavaServer Faces
We had Tapestry, Swinglets, and other technology which wasn't looked at as closely as it should have been.
In JDO 2.0, we have a chance to LEARN from our experiences with JDO 1.0. Vendors have real users, with real needs. Users are on the expert group too.... asking for their pain points to be resolved. Hopefully this will make the outcome more of a success than anything else. But then again, what will EJB 3 do to JDO? We will find out more next week!
So, lets standardize after the fact... follow proven technology, which is working in the real world. Then we will have great standards, that do their job.
Previous Topic
