OK, another question. I work on a large product that has evolved over the last five years, and has lived through the release of Java 1.3 and 1.4. We specify a required level of Java, but the decision to adopt a new major level of Java has very little to do with new language features - unless new language features offer significant advantages when implementing a new feature (or, exceptionally, reasons to reimplement an old one) they will not be considered a reason for us to change. The main reason to upgrade is if some useful package is being brought into J2SE SDK (JNDI in the case of v1.3, JSSE/JAAS/JCE in the case of v1.4), which makes things easier from a service point of view.
If relevant features weren't introduced then we wouldn't move until the existing Java level went out of service (i.e. if it wasn't for JSSE we would still pre-req Java 1.3.1, which seems to still be supported by Sun). Once we have moved to a new major level of Java, we will start to use new language features in areas of new development (old code is unchanged unless it uses newly-deprecated methods; why risk breaking something that works?).
In the current climate, new development projects are uncommon - the vast majority of Java development will be focussed on existing projects - do you think that Java 1.5 introduces any new features (apart from the headline language changes) that would be a reason to upgrade from Java 1.4? Do you think Java 1.5 uptake is likely to be widespread?