Re: Combining latin small letters with diacritics

From: Philippe Verdy <>
Date: Mon, 5 Mar 2012 22:09:00 +0100

Le 5 mars 2012 21:32, Ken Whistler <> a écrit :
> On 3/5/2012 11:56 AM, Philippe Verdy wrote:
>> But the document cited by Denis gives a much more productive way that
>> allows stacking any kind of letters with its diacritics. There won't
>> be enough space in the BMP for such Latin supplements.
> If such generic use occurs,  then it basically requires rich text or other
> markup.
> That would represent an introduction of general mechanisms comparable to
> techniques
> used in mathematics. But in the cited document I don't actually see any
> usage that
> would be problematical. Do you?

For example I see a LATIN SMALL LETTER WITH MACRON, stacked above
other Latin vowels. And sometimes with other diacritics of their own.

> The characters in the DAM are a particular set used in German dialectology,
> and do not require the introduction of generic mechanisms.

>> The alternative could be to encode only a single combining diacritic
>> (or format control) meaning "Stack the next grapheme cluster above the
>> previous one", to override the normal side-by-side presentation of
>> successive grapheme clusters. and eventually a similar special
>> diacritic or control to create a stack below.
> That would be another way of sneaking a glyph description language into the
> character encoding. I don't think it should happen.

Not necessarily, this can be defined as a new set of properties
defining how those diacritics can be infered from a stacking property
(<above>, <below>, <small>...) and the associated characters. A new
kind of decomposition mapping, that will also help text renderers to
display them, without necessarily having to add many mappings from the
existing Unicode-encoded diacritics.

And this would pave the way to encode later such mechanism directly in
Unicode (the mapping above offering the compatibility), without having
to encode every complex Latin letter grapheme clusters as new
Received on Mon Mar 05 2012 - 15:11:37 CST

This archive was generated by hypermail 2.2.0 : Mon Mar 05 2012 - 15:11:38 CST