They keep using that word, but I do not think it means…
…what they think it means. The feature they’re describing isn’t “ligatures”.
Ligatures in English and other European languages are basically stylistic choices, not a requirement of the language. Other European languages might use what look like ligatures to English speakers, but are actually letters in their own right: To Danes, “æ” is just as much a letter as “w” is to English-speakers, and the French “œ” represents a different sound to the two letters “oe”, and “ß” isn’t interchangeable with “ss” in German (Swiss notwithstanding). Every other joining of letters is done to make text look “interesting” or to solve kerning problems, like ff, fl, fb, ffi, and fi in typefaces where the f glyph has a long overhang to right. Anyone who actually types the “fl” codepoint (U+FBO2) into a document is causing more harm that good - that code only exists for round-trip encoding compatibility with the original MacRoman 8-bit encoding from 1984 - it is not a character.
(As an aside, I have a pet hate of monospaced fonts that form ligatures, and I reserve a special place in hell for the ones that try to get all arts-and-crafts-movement with the punctuation combinations used by programming languages: after decades of training my eyes to find &&, <=, !=, +=, ++, ==, ?. and other digraphs in code, rearranging them into artsy shapes only hides them from view)
Unlike Latin scripts, in calligraphic scripts like Arabic, ligatures are essential to creating legible text, rather than something that looks like a jumble of individual letters thrown on a page.
But from the sound of it, and the examples used, this release isn’t talking about ligatures at all, but what’s called contextual rearrangement: the way that some writing systems lay out glyphs (the shapes that we in English call “letters”, “numbers” or “punctuation”) in a different order to the reading or storage order. This is common to pretty much all of the scripts used to write the native languages of India, Bangladesh and Sri Lanka.
Here's an example in Sinhala, as used in the Sri Lankan language of the same name: ඵ + ේ → ඵේ
If your browser rendered that right, you'll notice that the second character has “wrapped around” the first, with one part of it being rendered to the left, and the small diacritic mark from the second character (the bit that looks like a p) now sitting to the right.
These days, most software can display this sort of thing properly (it uses a finite-state automaton described in the text renderer and/or the font itself to choose the appropriate cluster glyph based on the input sequence), but for an editor, locating your cursor within one of these clusters as you move it “left” or “right” is a more complex problem.
Biting the hand that feeds IT
