Resolver
Not a lot to talk about today. The resolver is coming along reasonably well.

I've had to build Query objects from scratch, which in turn has meant that I've had to build a heap of other objects with them. However, the parameters are only defined by interface, and the javadoc isn't as helpful as it should be.

While I agree that parameters should be defined by interface, I had no idea what classes should be instantiated as parameters. Grepping the source for one interface name led to other interfaces which extend the first, and a couple of abstract classes for good measure. Grepping for these found even more interfaces and abstract classes. Eventually the required classes would show up, but occasionally I'd get lost, and have to go back up the inheritance tree again.

Building the Javadoc would help this, only I have a non-compiling build at the moment, so I couldn't. Given the time wasted I should have just got a new checkout specifically to build the javadoc. Funny how you don't see the obvious when you're working too close to it. Alternatively, I could just ask SR, but I'd end up doing that so often that it would be like him writing the code rather than myself.

Ideally we'd properly document the important parts of the class hierarchy, for use by developers who are working in that code for the first time (like me). However, that takes a lot of valuable time, and with all the pressures we're constantly under, there's no way we'd be allocated the time for something like that. Theoretically I could do it in my own time for Kowari. That's one of the advantages of it being open sourced. However, I have more interesting things to spend my own time on. :-)

Masters
Tomorrow I have an interview at UQ about research topics. Wish me luck!