A Question About the Use of Irregular UTF-8 Expressions as the Solution to Emphasis in Gemtext

This is a response to this post:

gemini://pandion.midnight.pub/emphasis.gmi

This is not a new idea. Someone, somewhere, has already shown they can construct complex mathematical diagrams using this technique. However, Pandion suggests we might use this method for bold and italics in Gemtext.

Background:

> So I was looking for a way to insert a non-breaking-space* on android, when I stumbled upon the app: 𝘪𝘳𝘳𝘦𝘨𝘶𝘭𝘢𝘳 𝘦𝘹𝘱𝘳e𝘴𝘴𝘪𝘰𝘯𝘴 𝘬𝘦𝘺𝘣𝘰𝘢𝘳𝘥.

> This opened a whole set of possibilities, like:

> 𝗕𝗼𝗹𝗱, 𝘪𝘵𝘢𝘭𝘪𝘤𝘴, ｗｉｄｅ, sᴍᴀʟʟ ᴄᴀᴘs, u͟n͟d͟e͟r͟l͟i͟n͟e͟d͟, ̶s̶̶t̶̶r̶̶i̶̶k̶̶e̶̶o̶̶u̶̶t̶, 𝔬𝔩𝔡 𝔈𝔫𝔤𝔩𝔦𝔰𝔥, 🄱🄾🅇🄴🅂,

> 🅱🅾🆇🅴🆂, 𝚖𝚘𝚗𝚘𝚜𝚙𝚊𝚌𝚎, 𝕔𝕦𝕥𝕠𝕦𝕥, 🇲​🇦​🇷​🇮​🇹​🇮​🇲​🇪​, ˢᵘᵖᵉʳᵗᶦⁿʸ

NOTE: The app is called "irregular expressions keyboard".

> Those are simply utf-8 symbols, that anyone can write, and – 𝘩𝘰𝘱𝘦𝘧𝘶𝘭𝘭𝘺 – every client can support.

> They don't require rendering, and don't depend on the client.

> They are actual 𝘳𝘢𝘸 𝘵𝘦𝘹𝘵, and not "rich text" or anything complicated.

Whether or not these characters render properly will depend upon the context. My system does not render these characters properly. In Lagrange, due to its built-in fonts, the text renders as the author expects. But in the terminal, it does not. Even Vimini, a graphical browser, does not render them correctly. When the technique fails, it fails horribly, leaving the text littered with inscrutible unicode boxes.

What determines whether these characters render properly or not?

That's my question to you.

Posted in: s/Gemini

🚀 blah_blah_blah

Apr 19 · 3 weeks ago

4 Comments ↓

🕹️ skyjake [mod...] · Apr 19 at 18:46:

This topic comes up every now and then. Here is some previous discussion:

— /s/Geminispace/3228

My stance on this:

> As a matter of principle, Unicode characters should be used according to their semantic meaning and not their visual appearance.

>

> The appearance may vary wildly on different devices and apps. Relying on any single visual appearance, the one that you're seeing, will cause some percentage of readers to be unable to decipher your message. The screen reader case is just the extreme example of this.

🚀 stack · Apr 19 at 19:07:

Yeah, using available weird characters in a font is often "invented". If you want your text not searchable or otherwise parseable (think accessibility for blind people, or translation), great.

🛰️ lufte · Apr 19 at 19:39:

The semantics of a character is not always well defined (read the fascinating story of RIGHT ANGLE WITH DOWNWARDS ZIGZAG ARROW at [1]) but I agree that one should consider they will not render equally (if they do at all) in all systems.

To your question of why they don't render everywhere, I know it's (at least) a combination of the available fonts (Unicode is constantly growing, and fonts need to catch up) and the capacity of the software to actually draw all symbols, even if the font has them. For example, Iced (the GUI toolkit I use for Vimini) didn't support this at all in its first versions (see [2]), and I'm not sure why it doesn't support *all* symbols yet.

— [1]

— [2]

☕️ Morgan · Apr 21 at 10:33:

I wrote about emphasis a few times on my capsule, I'm still pretty happy with <my> suggestion on the topic, but I <<don't>> think it's likely to ever go anywhere ;)