L2/14-200 Title: Don't use U+E01EF VARIATION SELECTOR-256 Source: Behdad Esfahbod Date: August 6, 2014 I hope the committee agrees that 255 variation selectors ought to be enough for all uses, ever. As such, I like to ask the committee to commit to not using U+E01EF VARIATION SELECTOR-256 for any encoded variation sequence. The reason I'm asking this is purely technical. I was adding support for variation selectors to some of the font access tools that I maintain, and I noticed that whereas previously font APIs exposed a font as a set of supported Unicode characters, now they should expose fonts as a set of supported Unicode characters plus their variations. It occurred to me naturally that I can use the high byte of a 32-bit integer to include the variation selector number if there's any, and use the lower three bytes for the Unicode character number. This works and is very easy to document / explain, except that it doesn't work for VARIATION SELECTOR-256. To put it another way, currently Unicode allows up to 257 variations of a character: one default, and 256 using the selector characters. I'm asking that this be limited to 256 instead, by dropping the last variation. Hope I explained this clearly. Thanks, -- behdad