+1 to simplifying and just dealing with cursors for now.

You can probably simplify domain notes of when pre-start

and post-end are valid by stating that, outside of the general

"String Cursors" section, pre-start is never valid, and post-end

is only valid as an end argument.

string-cursor-ref and string-ref/cursor are redundant.

The string-split/cursors signature

  (string-split/cursors string [separator [ limit ] ] start end)

suggests start/end are always required, but sep/limit are

optional.  Anything other than

would be confusing.

Splitting on predicate is useful and in my experience more

common than substring.  The opposite (tokenizing) and

both operations together (partitioning) can also be useful.

string-pad-both/cursors is well defined and useful for centering.

string-compress/cursors is a surprising name, since it has

nothing to do with general compression.  There's no ambiguity

in string-remove-consecutive-duplicate-chars/cursors, but if you

want brevity then "collapse" or even "squash" may be better than

"compress".

I think the missing string searching and string-join are essential.

Other useful procedures: prefix?/suffix?/mismatch, string-chop

(chop into fixed-length chunks), string-replicate (xsubstring),

string-replace (from/to substrings), string-translate (from/to chars).

-- 

Alex