I must say that I strongly dislike the semantics of FLUID-LET, because
it just can't be used in a multithreaded environment.  Now that there
is a draft SRFI for a thread system, perhaps some efforts should be
put in specifying a dynamic environment binding construct that works
well with it.  I hope FLUID-LET is dropped before too many pieces of
code (especially libraries) are written that depend on it, because it
won't be possible to reuse this code in a multithreaded environment.

Marc