lianghai
commented
4 years ago See also John Hudson’s clarification (https://twitter.com/TiroTypeworks/status/1237478665736224768):
It may help people understand the issue to look at it from a higher level. Bindi is orthographically the correct nasalisation sign to use with this vowel sign, but glyph processing is not the correct layer at which to perform such a correction. /1
Ideally, this would be a text input issue only, i.e. users following normal orthography can be expected to use bindi in this situation. If tippi is used instead, then this is something that a spellchecker should catch and suggest a correction to bindi. /2
Performing a shape substitution at the glyph processing layer to make an orthographic connection is problematic for three reasons. a) It introduces a confusable presentation of the encoding, which has potential security implications. /3
b) It obscures an orthographic error at the encoding level, which means someone visually proofreading the text won't catch the error. c) It prevents display of non-standard orthography, e.g. in a manuscript transcription, in which tippi may be intended. /END
Haven’t done a thorough review but it seems that at least vowel signs ii/e/ai/o/au are affected.
Baloo and Baloo 2 don’t seem to be affected.
(Gurmukhi’s behavior of having complimentary distribution of tippi and bindi on vowel letters/signs is represented on the encoding level, and the user experience may be optimized in contextual keyboard layouts. Fonts need to reflect what is exactly in the text string.)