I skimmed the revised spec, didn't see obvious problems.  I haven't
had a chance to actually implement and try it in Gauche yet, but
I hope soon I can.

The only comment so far is that the rather lengthy rational section;
I understand you have lots to say and they're all useful, but
the first time readers may wonder why you need so much rationals,
before she actually read the specification.  It occurred to me that
it might help if examples are presented before rationals.
(But the reader can click the link to jump to the specfication anyways,
so the order may not be important so much...)

A general comment---I don't know if future lispers will use this
alternative notation a lot, but when I provide a capability
to extend applications written in Gauche, customers sometimes
asked more traditional notation than bare S-exprs, and I needed
to hack up some crude syntax every time.  I wished if
there were an off-the-shelf alternative syntax that maps nicely
to S-expr.  This srfi (plus sweet-expressions) may serve as a
good option.