References: Character Committee Report, 7/25/89, Proposals 2.2.1 and 2.4.3
Edit history: Version 1, 1/3/90 by Kim A. Barrett
The Character Committee Report Proposal 2.2.1 (passed 3/89) requires
implementations to support and document a STANDARD character subrepertoire,
whose elements are the specified set of characters. This set of characters
corresponds exactly to those characters which are of type STANDARD-CHAR.
The Character Committee Report Proposal 2.4.3 (passed 6/89) states that
every character repertoire name is a type specifier.
Thus we have two atomic type specifiers for precisely the same thing. The
type STANDARD is equivelent to the type STANDARD-CHAR.
Change the name of the STANDARD character subrepertoire required by
Character Proposal 2.2.1 to be STANDARD-CHAR.
This removes the duplication.
Cost to Implementors:
Cost to Users:
Eliminates possible confusion when a person reading some code sees
(TYPEP foo 'STANDARD)
and wonders "STANDARD what? Transmission?"
Pitman and Barrett support this proposal (RENAME).
There was initally some concern that STANDARD-CHAR might not be a valid
repertoire name, but there are no restrictions placed on the names of
repertoires in any of the proposals in the Character Committee Report
dated 7/25/89. There is a footnote (#15) that constrains character
scripts and labels to only use Latin capitals A-Z, hyphen, and digits
0-9, which the name STANDARD-CHAR satisfies. Since this is a footnote,
it can be argued that it has no force anyway.
Unfortunately, this still doesn't remove STANDARD as a defined name
(i.e., exported symbol of the cl package) since it's used by CLOS for
what might be argued to be an equally ungeneric purpose. There's a
fair chance that somebody somewhere along the line is going to get
annoyed by the inter-package sharing that occurs due to this symbol
Additional comments on the write-up:
Moon (9 Jan 90):
I support this. I think it's only an accident of the process for
amending the character committee's proposals that the duplication
between STANDARD and STANDARD-CHAR was overlooked. Given that we
want to get rid of one of the two duplicate names, it's clear
that STANDARD-CHAR is a better name.