Email list hosting service & mailing list manager

strings draft Tom Lord (22 Jan 2004 05:11 UTC)
Re: strings draft Shiro Kawai (22 Jan 2004 09:46 UTC)
Re: strings draft Tom Lord (22 Jan 2004 17:45 UTC)
Re: strings draft Shiro Kawai (23 Jan 2004 05:03 UTC)
Re: strings draft Tom Lord (24 Jan 2004 00:45 UTC)
Re: strings draft Matthew Dempsky (23 Jan 2004 20:01 UTC)
Re: strings draft Shiro Kawai (24 Jan 2004 03:26 UTC)
Re: strings draft Tom Lord (24 Jan 2004 04:31 UTC)
Re: strings draft Shiro Kawai (24 Jan 2004 04:49 UTC)
Re: strings draft Tom Lord (24 Jan 2004 19:01 UTC)
Re: strings draft Shiro Kawai (24 Jan 2004 22:15 UTC)
Octet vs Char (Re: strings draft) Shiro Kawai (26 Jan 2004 09:58 UTC)
Strings, one last detail. bear (30 Jan 2004 21:12 UTC)
Re: Strings, one last detail. Shiro Kawai (30 Jan 2004 21:43 UTC)
Re: Strings, one last detail. Tom Lord (31 Jan 2004 00:27 UTC)
Re: Strings, one last detail. bear (31 Jan 2004 20:25 UTC)
Re: Strings, one last detail. Tom Lord (31 Jan 2004 20:56 UTC)
Re: Strings, one last detail. bear (01 Feb 2004 02:28 UTC)
Re: Strings, one last detail. Tom Lord (01 Feb 2004 02:58 UTC)
Re: Strings, one last detail. bear (01 Feb 2004 07:53 UTC)
Re: Octet vs Char (Re: strings draft) bear (26 Jan 2004 19:04 UTC)
Re: Octet vs Char (Re: strings draft) Matthew Dempsky (26 Jan 2004 13:13 UTC)
Re: Octet vs Char (Re: strings draft) Matthew Dempsky (26 Jan 2004 13:41 UTC)
Re: Octet vs Char Shiro Kawai (26 Jan 2004 23:38 UTC)
Re: Octet vs Char (Re: strings draft) Ken Dickey (26 Jan 2004 19:40 UTC)
Re: Octet vs Char Shiro Kawai (27 Jan 2004 05:10 UTC)
Re: Octet vs Char Tom Lord (27 Jan 2004 05:37 UTC)
Re: Octet vs Char bear (27 Jan 2004 08:35 UTC)
Re: Octet vs Char (Re: strings draft) bear (27 Jan 2004 08:32 UTC)
Re: Octet vs Char (Re: strings draft) Ken Dickey (27 Jan 2004 06:50 UTC)
Re: Octet vs Char (Re: strings draft) bear (27 Jan 2004 19:06 UTC)
Re: strings draft bear (22 Jan 2004 19:05 UTC)
Re: strings draft Tom Lord (23 Jan 2004 02:06 UTC)
READ-OCTET (Re: strings draft) Shiro Kawai (23 Jan 2004 06:00 UTC)
Re: strings draft bear (23 Jan 2004 07:04 UTC)
Re: strings draft bear (23 Jan 2004 07:20 UTC)
Re: strings draft Tom Lord (24 Jan 2004 00:15 UTC)
Re: strings draft Alex Shinn (26 Jan 2004 01:58 UTC)
Re: strings draft Tom Lord (26 Jan 2004 02:35 UTC)
Re: strings draft bear (26 Jan 2004 02:35 UTC)
Re: strings draft Tom Lord (26 Jan 2004 03:01 UTC)
Re: strings draft Alex Shinn (26 Jan 2004 03:00 UTC)
Re: strings draft Tom Lord (26 Jan 2004 03:27 UTC)
Re: strings draft Shiro Kawai (26 Jan 2004 04:57 UTC)
Re: strings draft Alex Shinn (26 Jan 2004 04:57 UTC)
Re: strings draft tb@xxxxxx (23 Jan 2004 18:48 UTC)
Re: strings draft bear (24 Jan 2004 02:21 UTC)
Re: strings draft tb@xxxxxx (23 Jan 2004 02:09 UTC)
Re: strings draft Tom Lord (23 Jan 2004 02:42 UTC)
Re: strings draft tb@xxxxxx (23 Jan 2004 02:44 UTC)
Re: strings draft Tom Lord (23 Jan 2004 03:07 UTC)
Re: strings draft tb@xxxxxx (23 Jan 2004 03:04 UTC)
Re: strings draft Tom Lord (23 Jan 2004 03:29 UTC)
Re: strings draft tb@xxxxxx (23 Jan 2004 03:42 UTC)
Re: strings draft Alex Shinn (23 Jan 2004 02:34 UTC)
Re: strings draft tb@xxxxxx (23 Jan 2004 02:42 UTC)
Re: strings draft Tom Lord (23 Jan 2004 03:02 UTC)
Re: strings draft Alex Shinn (23 Jan 2004 02:58 UTC)
Re: strings draft tb@xxxxxx (23 Jan 2004 03:13 UTC)
Re: strings draft Alex Shinn (23 Jan 2004 03:18 UTC)
Re: strings draft Bradd W. Szonye (23 Jan 2004 19:31 UTC)
Re: strings draft Alex Shinn (26 Jan 2004 02:21 UTC)
Re: strings draft Bradd W. Szonye (06 Feb 2004 23:30 UTC)
Re: strings draft Bradd W. Szonye (06 Feb 2004 23:33 UTC)
Re: strings draft Alex Shinn (09 Feb 2004 01:45 UTC)
specifying source encoding (Re: strings draft) Shiro Kawai (09 Feb 2004 02:51 UTC)
Re: strings draft Bradd W. Szonye (09 Feb 2004 03:39 UTC)
Re: strings draft tb@xxxxxx (23 Jan 2004 03:12 UTC)
Re: strings draft Alex Shinn (23 Jan 2004 03:28 UTC)
Re: strings draft tb@xxxxxx (23 Jan 2004 03:44 UTC)
Parsing Scheme [was Re: strings draft] Ken Dickey (23 Jan 2004 08:07 UTC)
Re: Parsing Scheme [was Re: strings draft] bear (23 Jan 2004 17:55 UTC)
Re: Parsing Scheme [was Re: strings draft] tb@xxxxxx (23 Jan 2004 18:50 UTC)
Re: Parsing Scheme [was Re: strings draft] Per Bothner (23 Jan 2004 18:56 UTC)
Re: Parsing Scheme [was Re: strings draft] Tom Lord (23 Jan 2004 20:39 UTC)
Re: Parsing Scheme [was Re: strings draft] Per Bothner (23 Jan 2004 20:57 UTC)
Re: Parsing Scheme [was Re: strings draft] Tom Lord (23 Jan 2004 21:57 UTC)
Re: Parsing Scheme [was Re: strings draft] Tom Lord (23 Jan 2004 20:20 UTC)
Re: Parsing Scheme [was Re: strings draft] tb@xxxxxx (23 Jan 2004 21:22 UTC)
Re: Parsing Scheme [was Re: strings draft] Tom Lord (23 Jan 2004 22:52 UTC)
Re: Parsing Scheme [was Re: strings draft] tb@xxxxxx (24 Jan 2004 06:48 UTC)
Re: Parsing Scheme [was Re: strings draft] Tom Lord (24 Jan 2004 18:55 UTC)
Re: Parsing Scheme [was Re: strings draft] tb@xxxxxx (24 Jan 2004 19:34 UTC)
Re: Parsing Scheme [was Re: strings draft] Tom Lord (24 Jan 2004 22:02 UTC)
Re: Parsing Scheme [was Re: strings draft] Ken Dickey (23 Jan 2004 12:53 UTC)
Re: Parsing Scheme [was Re: strings draft] Tom Lord (23 Jan 2004 23:35 UTC)
Re: Parsing Scheme [was Re: strings draft] Ken Dickey (24 Jan 2004 16:10 UTC)
Re: Parsing Scheme [was Re: strings draft] Tom Lord (25 Jan 2004 03:14 UTC)
Re: strings draft Matthew Dempsky (25 Jan 2004 00:00 UTC)
Re: strings draft Tom Lord (25 Jan 2004 07:29 UTC)
Re: strings draft Matthew Dempsky (26 Jan 2004 16:53 UTC)
Re: strings draft Tom Lord (27 Jan 2004 00:44 UTC)

Re: Octet vs Char (Re: strings draft) Ken Dickey 26 Jan 2004 19:40 UTC

On Monday 26 January 2004 02:13 pm, Matthew Dempsky wrote:
> From my understanding there are a few issues being discussed here
> (with possibly overlapping consequences) -- O(1) performance of
> strings and using strings for octet sequences.
>
> SRFI-4 defines homogenous numeric vectors which seem like they would
> suffice for implementing octet streams independant of strings and, if
> not, a few Scheme implementations (larceny, MIT/GNU-scheme) provide
> byte-vectors.
>
> As for O(1) performance, there's undoubtably a need for
> non-O(1)-string-like-heavy-weight-objects (herein called ropes) which
> some implementations choose to implement and call them strings (as
> permitted by r5rs).  The current proposal seems to urge
> implementations to provide these as a seperate data type.

Well color me dumb, but I don't see why getting O(1) is such a big deal.

Let's say an implementation reads UTF-8 chars.  A uni-string could be
represented as a tuple.  The required part is a byte-vector of characters
that fit into 8 bits.  If characters have 16 bit representation, use a
second, 16-bit vector and a map of which chars are the long ones.  If 16 bit,
index into the 2nd array, else into the first.  The 2nd array could be
shortened by using the byte-vector to hold indexes into the 16-bit array in
the mapped/unused bytevector slots.  In the "worst case" (storage wise), we
only need the 16-bit array.  The mapping could be a bit-vector or whatever
makes sense in a particular implementation.

It is true that there are more code paths in this scheme (sic), but each kind
of representation "knows that" and this still makes refs & sets ~ O(1).

What am I missing here?

Thanks,
-KenD