I also have to agree with the others that "bytestring" is confusing in
the absence of a distinct bytestring datatype.

"Bytevector representing a string" is not a simple concept, as the same
bytevector represents a different string in different character
encodings. If ASCII or UTF-8 is wanted specifically, that should be in
the name IMHO. RnRS has `string->utf8` which returns a bytevector. So
bytevector-representing-utf8 procedures could be called `utf8-foo` and such.

(bytestring<? bytevector1 bytevector2) could just as well be named
`bytevector<?` unless I misunderstood something. For case-insensitive
comparison of bytevectors as strings, a name like `utf8-ci<?` would be
logical.