Just thinking out loud while working on spacing and kerning. But it wouldn’t it be a nice feature, when G2 could show in some way if a pair of glyphs you want to kern is a plausible combination.
I was inspired by:
http://www.urtd.net/x/cod/combinations
Especially for combinations of accented characters.
All combinations are plausible. Just because somebody ran searches against dictionaries and thinks they’ve nailed down all important kerning pairs doesn’t meant they’re right.
I am well aware that all combinations are plausible, that is how I kern. But in some cases, smaller one time jobs for example, it might save some work.
The problem is also that, while a pair may be unplausible in words of the languages you cover, it may very well appear in an abbreviation.
For instance, German sharp s (ß) can only appear after vowels in German words. But then I saw an abbreviation, I think it was Rßft. for Rußfilter or something like that, and the R crashed into the ß. Duh.
However, one exception I can think of is the case where a letter only appears in one language, like Tbar, which only appears in Northern Sami. Any combination with a letter that does not appear in that language, like Q, may be left out. But, of course, you can never be 100% sure. Perhaps it is used for a new transliteration of a Non-Latin script, you know.
Maybe you are right 