Lying fonts – good or bad?
With Unicode and OpenType, there are specifications about how certain things should be done: particularly encodings and OpenType layout features. But some things are not as well or easily supported in applications, which leads to the temptation for font developers to “lie” about the encoding or alternate glyphs, in order to get something that works more easily. What specific kinds of lies are font developers tempted by, and is this lying a Good Thing or a Bad Thing?
I’ve been talking about this a lot lately in discussions on the public OpenType list and in the Typophile forums, and I thought I should put all my thoughts in one place.