>     I suggest allowing literal printable Latin-1 characters.
>
> I suggest instead allowing any character to be encoded as utf-8.

Because we're having this debate, I suggest sticking with the draft's
policy of only allowing printable ASCII :)

> Did you consider the variant of allowing strings embedded in u8vector
> literals?
>
> #u8("zero-terminated" 0)

Also a good alternative.

If we can assume widespread compile-time evaluation of
`bytevector-append` and `string->utf8` on constant values, it seems less
important which particular syntax is chosen.