Second day of the conference proper, but I have been here since Saturday, and the tutorials started on Sunday, so I don't know how to number them. I have been too busy to ven try to blog much -- is fun.
Martin Poschle and I talked for a while about how to enhance Torque and OJB to allow for better code generation. He is going to look at adding relation generation, both 1:N and M:N, to the OJB generator in Torque. I promised to help. Ack. In looking at how to do it I finally got to nose around Torque's workings and it is pretty straightforward. It shouldn't be terrible to try to do it. I also have had a lot of conversations with Jim Moore (log4j committer). He's a very bright guy and gets to work with systems way bigger than I do so he has a lot of good ideas for performance improvements -- and ideas on how to specify outer joins hints in the mapping for relationships for the sql generator. I need to look into that.
David Bau (of BEA and XMLBeans fame) is brilliant. XMLBeans is very powerful and has a lot of potential particularly for data object type things, and a very innefficient path between objects, relational databases, and xml. If he is typical of BEA, and XMLBeans is typical of the quality of their code, than there is a tiny chance that their app server is worth what they want you to pay for it =). When I was talking with David after his presentation Ted Leung very pointedly informed me that I must become a contributor. When?Ovidiu Predescu, of Cocoon and Google, gave an excellent session on Cocoon's FlowScript. I am not a huge JavaScript fan, but during Stefano Mazzachi's earlier Cocoon session the irc channel was full of Cocoon developers talking about what is going on -- and other languages for FlowScript are coming in the very near future (Ovidiu mentioned that Scheme was the initial language used).
So much more, but those are some initial highlights =)