Thanks for commenting, Faré!
This is the approach used by Gerbil Scheme, that is a monorepo of plenty of libraries on top of a Racket-like module system on top of Gambit Scheme (but has been retargeted in the past and could be retargeted again), and to a point used by my Gerbil-Utils which is a set of more experimental extensions to Gerbil, that I use a staging area before inclusion to Gerbil itself (when, in my copious free time, I can get the stuff documented).
Since you don't mention any problems with the monorepo approach I assume it has been a success. If that's the case, it's good to have validation.
A lot of things could be done to make Gerbil nicer, more portable, more coherent, etc.—but it's a great starting point.
Like Racket, Gerbil adds quite a bit to standard Scheme. How much of the current Gerbil libs would run on standard Scheme (R6RS and/or R7RS) and would Gerbil users be interested in collaborating on a standard Scheme library collection?
I have a strong interest in all the interfaces being portable to every reasonably featureful Scheme that wants them, and so far others feel similarly. The code itself can be cond-expand'ed as needed.
