From: Kenneth Whistler (kenw@sybase.com)
Date: Thu Mar 29 2007 - 17:27:49 CST
Philippe Verdy said:
> So use: <U+0304, U+030D> (COMBINING MACRON ABOVE, COMBINING VERTICAL LINE
> ABOVE).
I agree with Philippe that a sequence of existing diacritics (with
that being one of the more likely alternatives) would be sufficient
for simply being able to represent the text in question digitally.
It would give you a unique representation for the text, distinguished
from the letters that have only the macrons. With fallback representation,
using most existing fonts, you'd be able to distinguish the text
just fine, although the vertical line above wouldn't tend to
display optimally (connected to the macron). And even that could
be fixed by simply creating a font with appropriate ligatures for the
diacritic sequences.
But...
> There's absolutely no evidence that this is a unique diacritic with
> distinct semantic.
This I completely disagree with. The phonetic guides for the
books in question *are* the evidence for a unique diacritic
with a distinct semantic. The text even spells out exactly
what that "semantic" is for the system in question. That
the semantic is not well-founded from either a modern
phonological point of view or as a symbol for an English-based
pronunciation guide, has no bearing on the distinctness of
the written mark as a symbol, as a unit of written text
representation, and hence as a valid candidate for being
encoded as a character.
The fact that a symbol has a graphological history as a diacritic
modification of another preexisting symbol does not disqualify
the result from being treated as a character in its own right.
In this particular instance, I think a case can be made
for either approach. Encode the diacritic-modified macron
simply with a sequence of the existing encoded macron and
another existing mark; or encode a new combining mark for
the diacritic-modified macron (and do not canonically
decompose the resulting encoded character).
> And the second diacritic used on top of the macron is not
> really clear.
Agreed. I think the printer simply grabbed "apostrophe" punches
from a poor quality set of cases for what was already a
hacked up set of metal type, where the macrons, breves and
such on letters had been created by a similar process in
the first place. The overpunched diacritics sometimes look
like vertical lines, sometimes tapered vertical lines, sometimes
a bit more like acutes, and sometimes more like comma-shaped
apostrophes.
--Ken
This archive was generated by hypermail 2.1.5 : Thu Mar 29 2007 - 17:30:05 CST