The abridged version of my log:Day 1:
Exchanged some email with others on the whole supplies/drains relation. It's a process which might be able to be automated, but that isn't my department anymore. The next big item on my list seems to be expanding the Unknown-relation query to be able to find relations between two entities, allowing for a chain of different relations. We can either go with exploring well defined patterns of relations, like "(has_part)*, contains", or a more general search of all relations. This latter option begins to mirror the shortest-path problem on an unweighted graph, so a breadth first search would probably do the trick. I read up some to refresh my memory about shortest-path, Dijkstra, and BFS. I coded some to specifically look at particular relationship chains, and started working on the BFS.Day 2:
Finished coding the breadth first search option to look for the relationship between two entities. I allow it to look at up to classes four links away, but this number is easily reset. Preliminary testing looks like it's ok (I had to tinker with the database and insert some data in to get the test cases I wanted). I'm a bit at a loss at what to do next without the anatomy people in town, so I started writing javadoc's for my functions.Day 3:
Documented most of the code, and all of the important, guts-of-the-program functions in the approved, javadoc style. Had a group meeting today, at which Peter presented me the sketch of code allowing the user to query using a chain of relations instead of just one. This presents some interesting possibilities. We can use them both in queries and as a potential path to look down in Unknown-relation queries. So now I have to figure out how to input from the user a chain of relations, process it to send it off to Peter's function, then process the answer. That shouldn't be too bad--I just need to read up on modifying/making the GUI and different Swing components I'll need.Day 4:
Spent today hip-deep in GUI and Java Swing code and docs. I add code to put a new button on the program, but what that button does is the tricky part. I spent most of today digging for the best components to use. JFrame vs. JInternalFrame vs. JWindow vs. JPanel. And then we come to the trickier part of JDialogs and JOptionPanes, which almost but not really do what I want. It's that last mile of "almost but not really" that has lead me on quite a chase to figure out how. So I think I have the basic structure; I just need to put the rest of it into code.Day 5:
Wrote the code with boldness, thanks to all the research in docs I did yesterday. Tested the GUI side of things and debugged. I actually don't want a JOptionPane after all; I can use JDialog and customize the buttons. So the GUI looks fine; I just need to hook up Peter's function to my code and clean up the answer for display.