> So what’s the problem with programming ligatures?
> (1) They contradict Unicode. [Goes on to mention U+FB01 LATIN SMALL LIGATURE FI and how a ligature for U+003D EQUALS SIGN, U+003E GREATER-THAN SIGN is indistinguishable from U+21D2 RIGHTWARDS DOUBLE ARROW]
No. Nonono. Hell no, one might say. (I’m frankly disturbed to hear this said by a typography expert, because this means something is deeply wrong, perhaps with my understanding here.)
It was never a goal of Unicode to have a code point for every ligature a typographer might ever want. For example, there is a code point for FI and for LONG S T, but not for FJ or FT. Like other “presentation forms” (a thing that Unicode is explicitly not supposed to contain), these exist only for roundtrip compatibility with old encodings (some ancient Adobe-Latin variant maybe?) and are deprecated for any other use (although I’ll be damned if I can find a clear statement to that effect on the Unicode website, there are notes in other places[1]).
The second part holds water to some extent, but given the amount of visual ambiguity afforded by allowing a text to contain arbitrary Unicode (which we must if U+21D2 is at all a consideration), no programming font can solve it, and no programming font tries.
> (1) They contradict Unicode. [Goes on to mention U+FB01 LATIN SMALL LIGATURE FI and how a ligature for U+003D EQUALS SIGN, U+003E GREATER-THAN SIGN is indistinguishable from U+21D2 RIGHTWARDS DOUBLE ARROW]
No. Nonono. Hell no, one might say. (I’m frankly disturbed to hear this said by a typography expert, because this means something is deeply wrong, perhaps with my understanding here.)
It was never a goal of Unicode to have a code point for every ligature a typographer might ever want. For example, there is a code point for FI and for LONG S T, but not for FJ or FT. Like other “presentation forms” (a thing that Unicode is explicitly not supposed to contain), these exist only for roundtrip compatibility with old encodings (some ancient Adobe-Latin variant maybe?) and are deprecated for any other use (although I’ll be damned if I can find a clear statement to that effect on the Unicode website, there are notes in other places[1]).
The second part holds water to some extent, but given the amount of visual ambiguity afforded by allowing a text to contain arbitrary Unicode (which we must if U+21D2 is at all a consideration), no programming font can solve it, and no programming font tries.
[1]: https://twitter.com/fakeunicode/status/945017346858532864