dark subforum ideas thread: xxx edition

the wor8===D git has a har8===D g, as 8===Do give, gimlet, giggle and gift. Gilt and girl and gibbous and gig and gill too. When the nearly i8===Dentical wor8===Ds git an8===D gin have 8===Different pronunciations, you can't only rely on orthography; most soft gs are from Latin origins, while most hard gs come from other roots, meaning you can't really try to apply the rules to an acryonym anyway. it's not latinate or germanic, just a contextless neologistic initialism.

That sai8===D, it's not a giraffe-ical interchange format and "jiff" soun8===Ds ba8===D. there's not much case for either from a linguistic perspective, imo, so it really comes 8===Down to whether you respect the inventor's (8===Dumb) preferences. Personally, I think that if the wor8===D 8===Does not come from any kin8===D of recognize8===D language source, and the rule really only says that Latin-based words with a following i have a soft g, then there's no reason to apply those rules to an initialism.