Daniel Hegyi writes: > In one letter Paul K. writes: > >Why not use Java to add GUIs for these tasks to Emacs. Again, I think the > >opposition Emacs versus IDE is more perception than reality. > > > In another letter Paul K. writes: > >I think what you are suggesting is impractical. I don't see how you > >can provide a robust and efficient extension framework that runs in a > >separate process from Emacs itself. Interprocess communications > >and synchronization pose very great, perhaps insurmountable > >difficulties for such a framework. > > > I don't quite understand what you mean. Are you saying that it is a good > idea to do _GUI_ development for Emacs in Java, but it is a bad idea to do > much generic development for Emacs in Java? Or are you saying that any type > (i.e., not just GUI) and a large amount of Java development is certainly a > possibility but you don't support the notion of Java APIs for Emacs and > consider a better solution the use of the beanshell? >
I think it is practical to use Java to extend Emacs where not much interaction is required between Emacs and the extension, the wizards and the debugger are examples. It may even be possible to provide those extensions with a Java GUI. Pettar Maehlen and I have been discussing the possibility of creating a Java window for JDEbug to display local variable values when stepping through a program. Currently displaying local variables slows down stepping significantly because it involves sending significant amounts of text across the standard I/O channel to Emacs. A separate local variable window that lives in the debugger process instead of the Emacs process would eliminate this bottleneck. - Paul
