@aral How about also fixing the screen readers? I'm not arguing against not using UTF8 special characters while screen readers can't properly interpret what was written - but since consecutive characters typically form words, why not teach that to screen readers?