Re: Code point vs. scalar value

From: Asmus Freytag <>
Date: Wed, 18 Sep 2013 08:34:07 -0700

On 9/18/2013 2:42 AM, Philippe Verdy wrote:
> There are "scalar values" used in so many other unrelated domains
> (notably in mathematics, where a scalar value is an identifiable
> object that remains constant in relation with some operations and
> independant of its context, unlike functions, differential or
> aggragating operators...scalar values may be sacalr only in some bases
> of the numeric domain), that using the terms "code point" is certainly
> more specific, less ambiguous, and will avoid more confusions with
> these application domains (including for example with algorithms used
> to format integer, real or complex numbers into encoded text, as used
> in localisation librabries like ICU wirh additional CLDR data...)
I concur. Codepoint is the accepted way of referring to the units these
algorithms operate on. Further, the input to an algorithm includes
erroneous input. Unpaired surrogates do exist in real data, and will
have to be dealt with by the error-handling part of any real
implementation of an algorithm.

( Aside: "Scalar value" as used in Unicode essentially denotes a
sub-range of valid code poin values, but many alogorithms and
implementations have their own subranges for various purposes, yet don't
all coin a comparable term to them.)

Received on Wed Sep 18 2013 - 10:38:26 CDT

This archive was generated by hypermail 2.2.0 : Wed Sep 18 2013 - 10:38:28 CDT