Email list hosting service & mailing list manager

Re: Update, near finalization Per Bothner (13 Apr 2005 22:55 UTC)
(Previous discussion continued)
Re: Update, near finalization

Re: Update, near finalization Per Bothner 13 Apr 2005 22:55 UTC

bear wrote:
> Actually, I think that specifying string behavior in a document
> about arrays is a mistake.
>
> The operations we want to do on strings are in many cases
> fundamentally different from the operations that are efficient to do
> on arrays.

I agree with you, but I think we're stuck with the fact that a Scheme
"string" is a low-level indexable modifiable fixed-length array of
characters.  Now this is a totally useless data type [*], except that
it's close to the hardware (unless you're using Unicode) so it
*might* be useful for implementing a more useful data type ....
in which case it should be hidden in the implementation.

Still, it's what we have.  Now if you'd design and implement a
more useful higher-level data type,d \ then maybe we can deprecate
low-level strings.able ut fixeb

[*] Modifiable but fixed-length makes no sense - except it's easy to
implement.  Indexing of strings is also a semantically bogus
concept.  To clarify: a position in a string, in the emacs "mark"
sense, does make sense - but the value of a position as an integer
isn't meaningful.
--
	--Per Bothner
xxxxxx@bothner.com   http://per.bothner.com/