Flaws in the BBCode 1.0 specification

I have noticed that people use [sub] where something like [small] was meant. The Shiichan implementation of BBCode 1.0 doesn't cover [small], which is kind of annoying.

We should design the 1.1 specs, since this effort looks like abandoned^[1]. Post all your proposals here. Particularly, IIRC, someone requested some LISP-like features. A formal description of such improvement would be nice.

^[1] http://wakaba.c3.cx/shii/

Why oh why does lisp seem so much better?
(lambda (expert x) (u (i (b (o x)))))
(expert 'EXPERT 'PROGRAMMER)

This can't be interpreted at server side, so imagine a javascript interpreter that translates your lispus to whatever format shii-chan finds friendly.

Actually, yes. That's what anon should do. Write a lisp-to-bbcode javascript interpreter that provides all the tags as functions:
b, u, i, o, etc
and lambda. That would be so awesome. Then we get moot to use it, in all boards or just /prog/. Or we use a de facto greasemonkey script.
I'd do it but I absolutely hate javascript.