> At 16:50 -0400 2003-07-30, John Cowan wrote: > >Michael Everson scripsit: > > > >> >See the reference glyph for U+FB4B. One form looks like this with > >> >the dot above further to the left, the other like it with the dot a > >> >little further to the right. This glyph with the centred dot is a > >> >compromise between the two. > >> > >> A picture speaks a thousand words. > > > >These particular words combined with the picture in the U3.0 chart tell all. > > I see. This disunification tempts. I'd go to the bother of writing up > the proposal for adding this combining character if on further > discussion it appears the right thing to do.
and Ted Hopp posted the picture that spoke a thou..., err, well two words in Hebrew and two character names: > I've posted an image at http://www.zigzagworld.com/holams.gif of two words > that illustrate one publisher's typographic distinction between vav with > kholam khaser and kholam male. Given this evidence, I would agree with Michael that the least disruptive approach would probably be a disunification of the point *in this case*. The vav-kholam would be represented using the currently available sequence: <05D5, 05B9> (U+FB4B HEBREW LETTER VAV WITH HOLAM is canonically equivalent to this and should be interpreted the same, no matter what the current glyph in the charts happens to look like, by the way. The glyph can be updated to shift the holam dot further left, more as it is shown for U+05B9 itself.) The kholam male would then be represented using a distinct dot above: <05D5, 05XX> It is *not* advisable, in my opinion, to try to represent the kholam male in Unicode as <..., 05B9, 05D5> (where the existing holam character would formally apply to a preceding base character, but be rendered over the vav). This because graphologically the dot is rendered on the vav, whatever the underlying phonological and historical analysis, and Unicode rendering is generally going to be more straightforward if its principles are consistent about association of diacritics (and points) with base characters. Then the question arises regarding the identity of 05XX. It could, of course, be a new dot above: U+05BA HEBREW POINT HOLAM FOR MATER LECTIONIS (kholam male) or whatever you want to name it. But how about: U+05C4 HEBREW MARK UPPER DOT What the heck is *that* thing for, and how would it be distinguished if it isn't this holam? Note that U+05C4 does not participate in any decomposition, so that isn't an issue here. 05B9;HEBREW POINT HOLAM;Mn;19;NSM;;;;;N;;;;; 05C4;HEBREW MARK UPPER DOT;Mn;230;NSM;;;;;N;;;;; They only differ in combining class (and glyph). Both are given the Alphabetic property currently, by the way, for whatever that is worth in the discussion. --Ken P.S. I have always considered that dots will be the death of Unicode. ;-) Maybe someday someone will write the Comprehensive Guide to Dots in Unicode... Perhaps we should just encode a new block of combining dots: 60 combining dots arranged in 4 quadrants of 15 each at 6 degree intervals, plus another 36 interior overlaying dots in a 6x6 grid of positions, and then just tell everyone to pick the dot that suits them best for whatever their intended purpose may be...
