Re: ISO/IEC 10646 as Document Character Set

Glenn Adams (glenn@stonehand.com)
Fri, 5 May 95 17:35:49 EDT

Date: Sat, 06 May 1995 16:10:33 -0500
From: eric@spyglass.com (Eric W. Sink)

Opinion: changing the doc charset is NOT an editorial change. If we
make the doc charset 10646, then applications must support 10646 to
be conforming. This spec defines conformance, and the fact that
existing applications support a SUBSET of 10646 does not make it
better.

I would disagree for a number of reasons:

1. SGML doesn't say whether or how to interpret a dedicated data character.
2. 10646 doesn't say whether or how to interpret any character.
3. 10646 formally provides for the identification of subsets by means of
two methods known as "limited subsets" and "selected subsets"

We could specify:

1. 10646 is the doc charset

2. an HTML 2.0 application is expected to support those graphic characters
which comprise the selected subset of 10646 consisting of the collections:

1 BASIC LATIN 0020 - 007E
2 LATIN-1 SUPPLEMENT 00A0 - 00FF

3. by "support" is meant being capable of displaying a suitable rendition
of a graphic character.

Given the above, how would it not be an editorial change?

Regards,
Glenn