I'm working with GE's Centricity Electronic Medical Record system (Logician). To debug forms, I can get a trace of the MEL language that they generate and the values returned. Making things more interesting is that functions can fire off semi-randomly, depending on what's getting updated. A while ago, I created a form to make these logs more readable. I attempted to break it up so that the upper window shows the statement, and the bottom window shows the trace that the statement generates. To a degree, this works, but it rapidly gets overly complex.

For example, the code to generate a RTF chart note is traced one expression at a time, so each "execute/results" pair gets longer and longer -- except for the shorter evaluations in the middle. Also, when an observation is changed, it may trigger multiple expressions that use that observation. Basically, it's a series of nested series of expressions, flattened out sequentially.

The way I'm doing it now breaks it up so that it's slightly more readable (White Space Is Your Friend), but I'm not completely happy with it. Something that would make me happier would be to be able to show the nesting more obviously, but with two windows and a scrollbar, that's hard to do. I was thinking about incorporating a Treeview control, but it seems like that would make the navigation more difficult, and use up screen real estate that might be better used to view lots of text.

It's hard to explain exactly what the input looks like without posting some, but I get the impression that I'm not supposed to do that: most of the online support information is fairly well locked down to existing customers, including the support board and mailing list.

So, is this making any sort of sense to people? If so, what would you do about it? If not, you have my sympathy, as it doesn't make a lot of sense to me, either. :-)