Three words on Mozilla application integration: what a pain. I don't know who's idea it was to use these contents.rdf files, but there must be something easier. I ended up having to put everything in one directory (/content) for it to work.

... but it did, finally. I didn't check the RAM numbers, but it loads on its own without Mozilla (I'll post a screenshot). There's probably quite a bit of overhead for such a small app.

My opinion of Mozilla as an application platform? Current Mozilla development has very weak language support. You can only write the UI scripts in Javascript and the backend in C++/XPCOM. Other languages (notably Python on both the UI and backend sides) are in the works. The Java to COM bridge that was in development by Sun seems to have been abandoned.

Another disadvantage, and this applies to web browsing as well, is that Mozilla is a moving target. You never know what a user is going to have installed. It's fine to develop for Mozilla 1.0, but if the user has 1.2, can the two versions co-exist or will your code work on both? I had some problems using examples (especially the RDF files and integration) out of the book (published in Sept, 2002) because they were already outdated. Unless you are in the thick of things, you never know how your application can break. This is not a good thing.

So for a spike, this was a pretty good one. I found out that Mozilla platform development isn't all that great yet. It has potential though, especially XPFE (the XUL, Javascript and CSS combination), which made the UI so much easier to layout (granted I have experience with DHTML). I would really like to see a way to use XPFE with Java in the near future. That seems like a great cross platform RAD solution.

Like HTML, a WYSIWYG XUL editor might be more of a hinderance than a convenience, but it's likely one will be developed anyway. One integrated into an IDE would be pretty nice.