I've just verified that the implementations of unquote and unquote-splicing in Guile do not conform to R6RS. Perhaps this is intentional, but it would be a shame because the change in R6RS is certainly one for the better. Scheme's quasiquotation is defined very well (it's even used as a reference in Common Lisp's backquote definition) but ... it did have a major drawback.
unquote (and unquote-splicing as well) cannot be applied to multiple expressions. This messes things up a little, especially when you're dealing with nested quasiquotations. Anyway, you can't do things like this (like you can in CL and R6RS):
(let ((name 'foo))
`((unquote name name name)))
=> (foo foo foo)
But the latter example is nothing. You can get around it in a second. There are however MUCH worse-case scenarios which, quite possibly, cannot be worked around (or at most with ugly solutions).
The current rules for evaluation are (from R6RS):
If an (unquote <expression> ...) form appears inside a <qq template>, however, the <expression>s are evaluated (“unquoted”) and their results are inserted into the structure instead of the unquote form.
If an (unquote-splicing <expression> ...) form appears inside a <qq template>, then the <expression>s must evaluate to lists; the opening and closing parentheses of the lists are then “stripped away” and the elements of the lists are inserted in place of the unquote-splicing form.
Any unquote-splicing or multi-operand unquote form must appear only within a list or vector <qq template>.
That just about says it all. Alan Bawden discussed the previous flaw in Scheme. Now it's cured. Hope to see the new rules in Guile too.