From: Kenneth Whistler (kenw@sybase.com)
Date: Tue Sep 06 2005 - 13:36:50 CDT
Marnen Laibow-Koser indicated:
> 1. If the reference glyph for MULTI-REST is changed, I fear that it
> will break existing fonts and documents. I hate to say it, but it
> might be better to just live with an improperly named character
> (perhaps MULTI-REST = breve rest, H-SHAPED MULTI-REST = multi-rest?).
I concur with this assessment.
The history of the musical symbols proposal makes it clear that
*intent* of the encoding all along was for what is now encoded
as U+1D13A to represent the double rest. It was clearly in a set
with the whole rest, half rest, quarter rest, etc. Its *identity*
has been established by that history and the now long-published
glyph for the character. Attempting to change that would be a mistake
at this point.
So the name is wrong, and needs to be added to the growing list
of incorrect character names. The correct way to handle this
is by annotation and acknowledgement of another name problem.
Changing the glyph would create a *bigger* problem than the problem
the defect report is attempting to fix. This is another instance
where a legalistic interpretation of the standard's claim that
names are normative (and cannot be changed) and that glyphs are
representative runs afoul of the deeper principle that it is
the *identity* of the encoded character which must be maintained
stable.
As for the requested additional symbol, U+1D1XY MUSICAL SYMBOL
QUADRUPLE WHOLE-REST, I don't see a need for encoding that. As
one of the examples indicates, a quadruple whole-rest is
"two double whole-rests joined as one". Given that all of these
symbols exist in the context of an implied higher level protocol
that does complex musical layout and formatting, it seems to
me that a sequence of the double whole-rests would be sufficient
for representation.
The other rationale for encoding -- citation as a musical symbol in
isolation in running text -- would seem to fail in the case of
the quadruple rest. It is not at all uncommon to see double whole-rests
cited this way (as shown even in the examples in the proposal), but
the citations of the quadruple whole-rests are all in staff layouts,
because they don't make much sense cited in isolation.
> 2. While it would be good to have a code point for the the H-shaped
> multi-rest, I can't think of too much actual use for it. Unlike the
> other rest symbols, the H-shaped rest shrinks or expands as necessary
> for justification, which makes a non-typographical implementation
> more likely.
The case for a symbol for an H-shaped multi-rest symbol is better,
I think. It is cited in-line in the example as a symbol. But it
should definitely be a newly encoded symbol, rather than a
reinterpretation of the existing U+1D13A.
For use in musical scoring, it could be used as a base symbol
associated with a start and stop anchor. But before suggesting
that that be the expected behavior, it would be best to investigate
existing scoring systems to see how they implement multi-rests.
It may well be the case that they simply use the existing
U+1D13A with a start and a stop anchor -- in which case even
the existing name for U+1D13A would not be entirely misleading.
--Ken
This archive was generated by hypermail 2.1.5 : Tue Sep 06 2005 - 13:37:54 CDT