A mutable string is clearly a location (object).
To understand the general case, we need to distinguish between length-immutability and content-immutability. These go together in SRFI 140, but in R[567] strings can be (and most of them are) length-immutable but content-mutable.
A length-mutable string is a location; a length-immutable string is not a location. Content-mutable strings contain locations, one per character; content-immutable strings don't contain any locations. By the same token, pairs/vectors are not locations; mutable pairs/vectors contain two/N locations; immutable pairs/vectors don't contain any locations.
If srfi-118 or srfi-140
is supported then the value of a mutable string can be empty, but it is
distinct from the empty string literals.
Yes, that makes sense.
According to my reading of r7rs,
no mutable string is eqv? to any other mutable string *or* any immutable
string, including "".
In R[567]RS:
* all strings are length-immutable
* all non-empty non-literal strings are content-mutable
* all non-empty literal strings may be content-mutable or content-immutable
* all empty strings are content-immutable
* all non-empty non-literal strings are not eqv?
* all non-empty literal strings may or may not be eqv?
* empty strings may or may not be eqv?