From: Asmus Freytag (email@example.com)
Date: Wed Apr 06 2011 - 17:46:59 CDT
On 4/6/2011 2:56 PM, Leo Broukhis wrote:
> On Wed, Apr 6, 2011 at 2:37 PM, Asmus Freytag<firstname.lastname@example.org> wrote:
>> On 4/6/2011 2:16 PM, Leo Broukhis wrote:
>>> Combining up tack above It can be represented by
>>> U+0304 COMBINING MACRON U+030D COMBINING VERTICAL LINE ABOVE
>> The question is not whether you can "build up" something that looks like the
>> proposed character, but whether the uptack "is" nothing but a combination of
>> these two.
> In the proposal, C.9 Can any of the proposed characters be encoded
> using a composed character sequence of either
> existing characters or other proposed characters? says "No".
OK, that's the take of the proposer.
Please note that the question does not read whether the character looks
like a character sequence, but whether it can be encoded as such. Small,
but telling difference.
> I've demonstrated that it is not true.
> There is no mention of this possibility in the proposal nor any
> justification as to why the proposed character must be represented by
> a single character.
Just because an uptack "looks like" a macron with a vertical bar,
doesn't mean it "is identical" to the character sequence. Just pointing
out that one can generate a shape like an uptack with this sequence is
not the same as proving the proposal author wrong in this instance.
At issue is not a glyphic analysis, but a character analysis.
If the notation the character is used in where to have a consistent
function for macron, and a consistent function for vertical bar, and the
function of the uptack were more or less that of a combination of the
two, then it might make more sense to use the character sequence.
However, Unicode doesn't use some character sequences, even though they
would seem possible. For example the solidus overlay is not used to
decompose the Danish ø. In general, Unicode is cautions on composition
where the combining character would have a glyph that attaches. This
would be the case here.
>> I don't know anything about the proposed character, but somehow I doubt that
>> it's fundamentally a composite.
> FWIW, a named character sequence can be added for it.
Too early to consider and whether you name a sequence or not, is just
window dressing, anyway
>> Could be wrong, though.
> Under magnification, it looks like a true tack rather than a
> combination of a macron and a vertical line. Is that a satisfying
> condition for encoding?
Well, it means that your "proof" of a possible decomposition has gotten
that much weaker.
This archive was generated by hypermail 2.1.5 : Wed Apr 06 2011 - 17:48:26 CDT