In a thread "The Imperative strikes back?" on comp.lang.functional a
couple of weeks ago, Brian Rogoff <[EMAIL PROTECTED]> mentioned:

> [..]  I really wish that
> the basic string handling capabilities in some FP languages were a bit
> better thought out. For example, I think in Haskell strings are lists
> of characters. Any Haskell folk care to provide a rationale for that
> decision? If I were going to restrict myself to one kind of built-in
> string type, I'd have made it an array of characters, or better, gone
> with subsequence references. Why? My experience is that strings behave
> at least like bidirectional conatiners, really like random access
> containers.  Subsequence references capture that behavior.

Other posters on the list mentioned the Standard ML Basis Library's
SUBSTRING signature (see
http://cm.bell-labs.com/cm/cs/what/smlnj/doc/basis/pages/substring.html)
.
This provides operations on a datatype "Substring" consisting of an
underlying string, a start index, and a length.  The idea is that the
usual substring operations of span, take, drop, etc., can be performed
without expensive copying.  Operations to convert between String and
Substring are provided, but these need only be performed when string
processing is complete; a more efficient internal representation (say
GHC's immutable byte array plus two Int#s) can be used during
processing.

Does anyone else think this is a brilliant idea that should be
implemented?  I float the idea in case (1) someone else is already
doing this, or (2) someone else is interested in doing it.  It
shouldn't be too hard, but I don't have time right now...

--KW 8-)




-- 
: Keith Wansbrough, MSc, BSc(Hons) (Auckland) ------------------------:
: PhD Student, Computer Laboratory, University of Cambridge, England. :
:  (and recently of the University of Glasgow, Scotland. [><] )       :
: Native of Antipodean Auckland, New Zealand: 174d47' E, 36d55' S.    :
: http://www.cl.cam.ac.uk/users/kw217/  mailto:[EMAIL PROTECTED]     :
:---------------------------------------------------------------------:




Reply via email to