Monday 8/12
Today I wrote list.jsp, which takes a bunch of hidden input values that I sent from main.jsp and makes a list of structures. I tried to make it ask the model database if there's a model for it, like the DSG does, but it seems that the DSG has problems with that. It uses:
(modeldb-find-node "structure")
This returns something for very few structures (one is 'ascending aorta'). It doesn't work for many structures that I know do have models.
So, either it'll check using (cgi-action) and see if it screws anything up, or I'll just list everything and try to add anything they ask for, notifying if something fails.
To do:
- Try using (cgi-action) to test the structures.
- Do the layout, the checkboxes, and the 'parent' checkboxes.
- Take passed values in.
Tuesday 8/13
The cgi-action method takes way too long and I think freezes, so for now I'll just list everything on the page and do the cgi-action check when they actually make the scene.
The list now looks decent, although the layout isn't always pretty when the attribute lists have significantly different sizes. I don't think that can be helped. There's some stuff that's hard-coded, because without a way to check terms, I have to prune out unwanted things by hand.
To do:
- Make a new method in AddStructure to call from Starter.addRequested(). It should check, make note in the error message if something isn't there, and add it if something is there. When all those calls are done, addRequested() should finish up the other calls and send the error, if there is one.
- Make the parent checkboxes work, either by finding a way to check all the children (preferable), or by passing their values by hidden inputs without checking.
Wednesday 8/14
The parent checkboxes now work, using JavaScript. When you check one, the children all get checked, and when you uncheck one, the children all get unchecked (at least in the browsers I'm using).
As far as I can tell, the preAdd() function works, as long as the graphics server is behaving. Occasionally, it seems to try to add things it shouldn't, and this corresponds with the process freezing. What I don't know for sure is which is causing which. My program could be the problem (in fact it's likely) but I don't know how. The only thing I can think of that might cause it is the cgi-action check.
So, when they go to make the scene, I'll make it use the (modeldb-find-node) command instead of going right to cgi-action, so that it'll hopefully stop freezing. This means you won't be able to add many things from the list, but you can get the idea, and you can add anything from the interface itself. The model database simply has to be fixed (and updated!) anyway.
To do:
- Make presentation.
- Make FME stuff accessible during presentation.
Thursday 8/15
The FME version is now working well enough to demonstrate. Details to be fixed in all versions:
- need multiple remove in JSP and FME
- shows part list when there are no parts in JSP and FME (and Swing applet, halfway)
- shouldn't be able to highlight a removed part with the selector bar
- invalid degrees should give an error window in applets
- ancestor/descendant calls aren't reliable (remove epicardium from heart, add epicardium, and heart becomes a part of epicardium)
- pop-up window layout is strange on some browsers
My presentation is just about ready, although it'll need some input from Jim, who'll be back on Monday. As long as I make sure Tomcat is running on this computer during the presentation, I can just link to 'orbit.biostr.washington.edu' to demonstrate it.
To do:
- Test and fix details.
- Make sure things will work in most browsers.
Friday 8/16
The details are pretty much fixed, as far as I can tell. The JSP stuff should work in any browser. I can't get the applets to work in Netscape 6, oddly enough. I suppose the only thing that really matters is making sure they'll work on the presentation computer, which I can check on Monday. The JSP ones are the finished product anyway.
To do:
- Get comments on presentation and on FME version.