This has bothered me for quite some time, and I haven't found an appropriate answer.
From what I understand, in Turkish the uppercase form for the dotted i is Idotaccent. Similarly the uppercase form of the dotlessi is a undotted I (a normal uppercase I in latin script). If this is true then building a c2sc and smcp should have a language specific variation to cover this. I have examined a number of OpenType fonts that have shipped with my Adobe apps, and I can't say that I have found any that do this. A normal lowercase dotted i transforms into a normal small cap I even when using a Turkish keyboard as imput.
How is this being handled by those of you developing OT?