New to Typophile? Accounts are free, and easy to set up.
We are having a discussion about which OT features are best suited for some substitutions for the Unifraktur blackletter fonts. Three of them are contextual: longs-replacement, r-rotunda replacement and an etc.-ligature (et represented by r-rotunda, as it is common with blackletter fonts). The fourth one shall replace Ä, Ö, Ü by Ae, Oe, Ue.
The ligature is quite easily done with [clig].
The long-s replacement shall not be on by default (personally, I don’t recommend automatic replacement on font level, but it can be useful on the web, where a fallback layout shouldn’t look too uncommon to the reader). It’s a quite simple replacement of s before lc-letters and ZWJ by ſ, so the writer has to know the orthographic rules, they can inhibit it using ZWNJ. Now I have seen such replacement implemented as [hist] or [ssXX], but both are recommended to be used with single substitution. I find [hist] is semantically wrong because the long s is not historic in blackletter but mandatory. As for [ssXX], there have been various threads where the argument goes, that for single character replacement [cvXX] should be used, which again is not contextual. Which problems might one face when using a contextual substitution with a non-contextual feature?
The r-rotunda replacement (after letters with a curved stroke on their right) is about the same problem as the long-s one. Just this one actually is historic.
The Umlaut replacement is a single substitution, that would go in [ssXX] or perhaps in [cvXX], probably both adequate?