For GIDs found in the subs coverage table, the application passes a GID to the feature and gets back a new GID. Note: This is a change of semantic value. Besides the original character codes, the application should store the code for the new character.
That exception is there because it’s good to substitute characters in this situation or would it be better to have superior and inferior characters as well as .subs and .sups? What about ordfeminine and ordmasculine mentioned in another topic, they are normally entered from the keyboard but i also want the .ordn feature – is adding something like ordfeminine.ordn an overkill?
In any font with both ordinals and a/o.superior, where those forms are
– In ‘ordn’:
sub a by ordfeminine;
sub o by ordmasculine;
– In ‘sups’:
sub a by a.superior;
sub o by o.superior;
Bearing in mind that the information is ten years old, it may have changed again. The entire post about it is here:
I jut found this unfinished post:
There are two options to name the superiors. ‘one.sups’ or ‘onesuperior’. The latter has a Unicode. The difference is, if you use the Unicode you have the superscript numbers encoded in plain text and that is semantically better an saver when copy pasting and such.
Thomas Phinney wrote in the OpenType mailing list about a decade ago:
“Seeing David Lemon’s post on subscript handling, it occurs to me that I should point out that I just blogged yesterday about a small change in our handling of ordinals, with regards to the Spanish/Portuguese ordfeminine and ordmasculine characters no longer be accessed by the ‘ordn’ feature.”
Writers of those languages are used to keyboarding the characters directly, anyway. Also, with the many fonts that have this substitution and no other ordinalized (sic) letters, having only these two letters affected by the ordinals feature confuses the heck out the users of all the other western languages.