It's been a good bit of debugging, and I've learned a lot! So far, I've discovered a lot about regular expressions (don't fear the geekery) because IE can't handle XML very well, which means switching my nice tidy XML responses from the server back into plain text for the AJAX to play with. Quite a variety of search-and-replacing came into play, especially to get rid of the white space (carriage returns and the like--carriage returns will be an upcoming blog: since when did anything return, and what's a carriage?) that appears here and there in web pages. In short, the XML tools led to nice elegant solutions, but the incomplete implementation of XML tools require icky moving back and forth between XML and plain text.

I've also learned that I have much to learn about error handling--giving useful info when something goes wrong. One might have thought that my years primarily as a teacher would have developed that, but only until one realizes that predicting student wanderings and guiding them back toward the fold is very differnt from predicting all the errors that code could produce. I think the heart of the matter is that with students the responsive approach is very helpful so that students can see and learn from their own work, whereas with some code much more proactive predictions of what could go wrong are called for. I don't like to be overly proactive with my students for fear of stifling something brilliant before it has a chance. With planning a new piece of code, exactly the opposite needs to happen, and I think that would speed development of new applications. Happily, that would bring me full circle back to making new things available the classroom for students to play with.

The upshot, I think that most of the recent bugs are fixed (at least in Firefox--haven't had time to go to Explorer, and it's gettin' kinda late.