Le 6 mars 2012 23:59, Asmus Freytag <[email protected]> a écrit :
> Usually, stacked accents do not get smaller, but generally just change > position, so this would have to be an exception to general stacked accent > layout. > > This is false. There's been a lot of cases were accents placed above capitals, for example, had to be reduced in height. This is what is happening here, to avoid increasing the line-height too much, or simply due to physical constraints on the metal to fermly hold the character types. The stacked letters are made smaller for the same reason : the whole (base letter, plus all diacritics) have to fit in a restricted height. Sometimes, it is the base letter that will be shrinked to make it fit the stacked diacritic(s). This occurs frequently with tilde, circumflex, caron, breve... When these are accents (acute, grave) their shape may be shanged to be more slanted or nearly flat. They may also ligate with the base letter (this has heppened so frequently for the upper ring, that now it is the standard way of showing it over capitals whose midle at the top as a stroke and leaves no gap for fitting the diacritic inside the left hole, e.g. A or O). It happened so frequently in polytonic Greek that most diacritics normally above have been moved from the upper center to the upper left as a standard (otherwise, it would be impossible to make the distiction between a spirit and an accent) that remains even today the modern script. Now I accept the fact that there may be diacritics attached below an upper letter, that merits a separate encoding: normally for cedilla goes below the base consonnant, but why couldn't there be an alternate CEDILLA above encoded ? we would encode it before the upper letter holding the cedilla). Consideration such as "logical binding" of diacritics with each other is not relevant here, it is an interpretation, and there's no visual distinction (the relative size of diacritics is not a good sign).
