I fantasize about a world in which Unicode standardized a 16x16 and 32x32 bitmap format. Early emoji were designed for display at low resolutions like those. Your phone could send a character sequence that decodes to a PNG image or something and then all emoji would be accurately sent and displayed as intended. If you wanted efficient support for the original emoji baked into Japanese cell phones you could define specific code points to be semantically equivalent to the sequence encoding the exact legacy bitmap.