Using the utilities package given below I’ve now managed to generate all of the implementation beans for my data model. I created a template that creates the guts of the code and then iterates through the columns using a different template to put in the getter/setter methods. The next job is to set up the many:many relationships and my own tagging system (that allows you to tag rows with extra data). Once this is done the rest should be easy, just change the template and the output target and leave the rest of the code alone.

Bean Managed vs Container Managed Persistence

I’m beginning to think I made a mistake here, going for BMP, because I can’t see how you can have a transaction across more than one bean method call with BMP. The only way I can conceive it is that you get some kind of transaction ID, which reserves you a database connection, and then pass that through to the beans. Then you’ll have to set up a getConnection() method that you pass the transaction ID to so you have the same one throughout. This would allow you to rollback the whole transaction if you had to. How then do you manage fail over between machines? I need to do some more reading I think. My system is low volume so it should be OK. I think I’ll need to implement some kind of optimistic locking strategy (cue Bee Gees Stradegy, when things go wrong and you can’t go on, Stradegy – sorry, that always happens to me whenever someone says that word, I think it’s all the years working for big corp IT, raddles the brain).

BT Voyager modem

Yes it’s still for sale. Someone left a message here but no email address. Contact me direct on francis (d0t) fish (aT) macmail (d0t)com if you want to buy it.

Regards all.