# [plt-scheme] Please help test version 359.100

>* Dave and I argue that an implementation can make up a new random value
*>* -- as long as it is a Scheme value -- and use it wherever the
*>* semanticssays unspecified.
*
Whatever we end up inferring from R5RS, the moral is that the word
"unspecified" is terribly ambiguous and should probably be avoided. In
particular, it does not tell you the scope over which the
lack-of-specification is quantified.
IOW, "unspecified value" could mean
forall Scheme implementations :
exists a value V called "unspecified value" :
forall applications of for-each :
the result is V
and
forall applications of string-set!
the result is V
and ...
or it could mean
forall Scheme implementations :
exists a value V called "unspecified value" :
forall applications of for-each :
the result is V
and
exists a value V called "unspecified value" :
forall applications of string-set!
the result is V
and ...
or it could mean
forall Scheme implementations :
forall applications of for-each :
exists a value V called "unspecified value" :
the result is V
and
forall applications of string-set!
exists a value V called "unspecified value" :
the result is V
and ...
Dave