Looking at the website I think I want to use struts for my project. There’s also a template-based technology, which is new to me, called Velocity, which looks good too. My ideal is to start from a description of the data model, with some tags for actions, in my XML document and then generate out the data entry subsystem complete. In order to do this I think I need to take a typical entity and then build the whole thing by hand.

So next tasks:

Generate DM and run script in Postgres (ironing the bugs out on the way).

Take an entity and “do” it, tagging subsystem, lookups and all.

Turn this into a template.

Write some code to take the XML and apply (via script or somesuch) the template across the whole DM.

It dosen’t matter if it’s rough around the edges, just as long as I don’t have to repeat code myself. Remember a good programmer is lazy. If you are doing something repetetive and can’t get the machine to do it then start again. I will spend a lot of time on editor macros because then you can repeat the boring shit over and over once you’ve done it once. That’s why I like bash and awk.

Still haven’t decided about EJB’s tho’, but they are a marketable skill.