On day 2 of JavaOne 2008 conference, Emmanuel Bernard talked about Bean Validation framework (JSR 303). Emmanuel said that a typical JEE application requires the constraints applied in database, data access, business, and presentation layers and the constraint implementation logic is usually duplicated in all these layers. The goal of JSR 303 spec is to provide a uniform and layer agnostic way to express and implement a constraint. It also provides a Constraint Repository API to expose the list of constraints for a specific JavaBean. The constraints are defined in the bean (domain object) so the validation logic is close to the domain model. They can be added to bean, field, getter method, or even a graph of objects. He explained the "cascaded constraint checking" feature used to validate an object graph.
Developers can also define their own constraints by writing a custom annotation and decorating it with @ConstraintValidator tag. Emmanuel said that JPA 2.0 framework could use the bean validation framework to generate database constraints based on the bean constraint annotations. Other frameworks like Web Beans, JSF 2.0 and AJAX libraries could also use it for the validation requirements. JSR 303 specification is based on Hibernate Validator framework. It will be released as part of Java EE 6 specification.
Earlier in the day during the keynote session, Thomas Kurian and Oracle developer team previewed the upcoming features of Oracle Fusion Middleware 11g which include Complex Event Processing and Business Activity Monitoring in their SOA Suite, WebCenter Suite for developing Enterprise 2.0 and Social computing applications. The developer team also demonstrated the grid computing support that BEA WebLogic and Coherence Grid products offer to the enterprise developers running their applications in a virtualized application server environment. They also showed JRockit administration console and how to configure the JVM to perform deterministic Garbage Collection.