Olivier Galibert <galibert(a)pobox.com> writes:
May I hear technical arguments _for_ modules for a change ?
OG.
Yes, here's one.
As part of my research, I am working on a multi-lingual program analysis
engine. I need a front-end. Actually, I need many front-ends: an
editor, a couple of graphical tools, may be a batch tool. So, what do I
do? Invent an extension engine? 10 years ago I would have done just
that, but now I a have more appealing alternative: I instrument the
analysis engine as a dynamic library with language bindings du jour.
You want to hack in perl? I give you perl bindings. Tcl? No problem.
Java? We have that. An editor? XEmacs! This is why *I* want modules
in XEmacs. I don't really care about TTY code not being loaded when I
am using X; or vice versa. (Although I admit that splitting things like
HTML parser would be a good engineering decision). I need
extensibility; I need even more extensibility than XEmacs lisp gives me,
and modules do just that.
Marat.