3 ms·
IMHO the main problem is that this ever-growing mass of random clip art needs to ship on every embedded system that renders user-generated text. Text rendering
by pavlov 7y ago
IMHO the main problem is that this ever-growing mass of random clip art needs to ship on every embedded system that renders user-generated text.
Text rendering has become a lot more complicated and memory-intensive. Glyphs used to be plain quadratic Bézier paths, but now there’s hundreds of full-color images — and if your drawings don’t look as pretty as Apple’s copyrighted emoji designs, someone will complain.
- int_19h 7y agoThe upside is that it significantly improved the adoption of Unicode in general. On desktop and mobile devices, this mostly meant more glyphs - "Unicode" fonts we had in common use in early 00s often wouldn't even have many common alphabets, much less math and science characters. I remember how some third party fonts specifically advertised full coverage of the BMP. But these days, the default system font will usually have close to full coverage. For embedded specifically, this forced Unicode into many systems that previously didn't have it, and sucked at non-ASCII (or non-Western European) text as a result. 10 years ago, I still had to use transliterated titles for my music files to get the in-dash system in my car to show them properly; not anymore. BTW, I don't think emojis have to be full-color images. On some platforms, sure. But e.g. on Win10, emojis are clearly vector images.