At least once a week, like clockwork, I see a post warning against abusing text characters from outside the standard alphabet (i.e maths notation, superscript) in posts. Some people use them for style reasons, especially places in that only support unformatted plaintext - like mastodon.
'Screen readers don't like them' is the basic reason to avoid using these characters. A screen reader encountering a username made of extreme 'zalgo' text might keep a text-to-speech unit busy for an hour as it explicitly describes every modifier and ligature.
But I have to wonder: are screen readers really so basic? Surely this kind of nonsense is so prevalent that a #TTS focused mastodon tool knows not to render usernames that use characters from the greek/math italic set? Or upon encountering a post filled with ascii art or obscure unicode characters that will take a long time to read, triggers some kind of heuristic to compress or otherwise skip? I understand that in the context of an maths article a sigma character represents 'summation', but when I see a capital 'Σ' at the start of a display name I instead understand it is either a greek user or just someone that wanted a 'fancy looking capital E'.
In an inclusive space we should prioritise accessibility to everyone, and if that means avoiding Fraktur unicode, fine. But I am wondering if #blind TTS users are really using software that has not adapted to the way other users interact with social media. Are there tools that handle Emokid2008s insistence on posting lyrics in bold italic unicode characters? Is GothicPrincess99's username a five minute description job because of the blackletter text she chose?
Hopefully my lack of knowledge in this area isn't too offensive. I just want to understand if the TTS social media experience is really so susceptible to non standard characters and more importantly: why is this the case??
(cc @GoemonIshikawa )
