Contrastive use of kratka and breve
Jukka K. Korpela
jkorpela at cs.tut.fi
Wed Jul 2 13:55:53 CDT 2014
2014-07-02 19:11, Leo Broukhis wrote:
> is an example of й and и + U+0306 COMBINING BREVE used contrastively
> (/j/ vs short /i/) thanks to a difference in typographic style of
> Cyrillic breve (kratka) and regular breve.
I can’t trell where the difference comes from, since this is a bitmap
image. But my hypothesis is that this has nothing to do with making a
difference in typographic style between kratka and breve. Rather, it is
a matter of mixing font: in one case, you have the normal letter й, and
in another case, you have the letter и and the combining breve *taken
from another font*.
> For me in Win7 using и + U+0306 results in a contrast,
This does not really depend on operating system; rather, in the
rendering system of the program being used and on the fonts used.
For example, using Word 2013, й and и U+0306 COMBINING BREVE look the
same for many fonts, but if you select a font that does not contain
U+0306, the results are odd. In some lucky cases, you might get a breve
symbol in the right place
Using и U+034F COMBINING GRAPHEME JOINER U+0306 COMBINING BREVE causes
different results. For many fonts, such as Calibri, the result is
different in the sense that the breve (kratka) has been displaced to the
right, and the result looks odd. Changing the font to Arial results in a
mess, since CGJ, which displayed visibly. Apparently there are problems
with CGJ in rendering software, even though CGJ should make no difference.
Using ZWNJ between the base character and the diacritic is more logical,
but it seems to yield similar results. The point is that whether or not
these control characters should prevent the rendering of a base letter
and a diacritic as a precomposed glyph, they tend to do that, but you
mostly won’t like the results.
> but given that и
> + U+0306 is a canonical decomposition of й and a renderer is allowed, if
> not encouraged, to use the glyph for й every time it sees и + U+0306,
> what is the right (portable) way to do that? Would и + ZWNJ + U+0306
> work? Should it?
As far as I can see, there is nothing that says that it (or any other
similar method) should work, but it may often “work”—in a manner you
might not like.
> I'd like to reply to
I think the idea of using CGJ is more wrong than the idea of using ZWNJ.
But neither works *properly*.
You can make a distinction between й and и with kratka (caron), at the
character level, and applications are allowed to render them
differently. But Unicode specifies no way to request for such rendering,
still less guarantee it.
More information about the Unicode