2008 March 31
Spent today working on reducing some sample FX64 data. Got a rough recipe together, and it seems to yield pretty good results. Writing the steps down makes a big difference in thinking about how to do things and making sure that you do a thorough job of looking at things.
The next step, I think, is to work with some more complicated data. Right now I’m just reducing a big blazing calibrator image, which doesn’t stress things much. I’d be interested in extending it to observations of more interesting sources. I think the extension will be straightforward, but it’d be good to check how things work. That sort of work isn’t too relevant to my project, though, so I may not spend too much time on it.
Another thing that I could do is unify my code a bit better. I have a lot of modules doing very similar things but with mutually incompatible implementations. Unfortunately, it’s difficult to get things talking to each other while also being flexible enough so that the different bits can do what they need to do. I have some vague ideas that I think are sensible, though, so I’d like to imagine that I can make my various modules at least play together a bit better. On the other hand, it’s definitely not a good idea to get too stuck on beautifying the framework, since that activity can consume unbounded amounts of time.
Overall, I do have a sense that the necessary tools are coming together. It’s never going to be perfect, but I can obtain the most of the numbers I want and visualize them how I want. Combine the ability to do that with a good recipe for reducing the data (which I am actively developing) and you’re about as good as you can ask for.