Why not have "no value" instead of "unspecified one" in mutations procedures (set!, set-car!, set-cdr!, vector-set!, vector-fill!)
By convention, the value returned by a mutation procedure is unspecified.
to
By convention, mutation procedures return no value.
PS: is there any implementation returning the value being stored ?
Thanks for bringing this to mind about MIT-Scheme, as at least one implementation returns a value it is indeed not wise to push such a change (at least for 'set!').
Duplicate of #68.
In MIT Scheme, set! returns the previous value of the variable.