3 ms·
Much of it is a tradeoff between features, generality and bloat. A good chunk of that gigabyte of RAM to render a webpage goes into stuff like vector font rend
by retrac 6y ago
Much of it is a tradeoff between features, generality and bloat. A good chunk of that gigabyte of RAM to render a webpage goes into stuff like vector font rendering so it can display nicely at any size on any display. And Unicode handling so it can display Chinese and Russian as well as English. I suppose both outline fonts and Unicode are bloaty in their own right; such allegations have been leveled against them before. Still, those sort of features come with pretty significant cost and are largely invisible until you need them.
- grishka 6y agoI am Russian, I know this whole encoding problem first-hand. I'm genuinely envious to people who can get by with just a single Latin-script keyboard layout. Unicode and vector fonts have been part of all operating systems for at least 15 years. Mac OS X 10.0 that came out in 2000 and ran on 128 MB of RAM and a slow by modern standards single-core G3? It supported both these things with ease. Windows has also had W versions of all API functions that accept strings, and supported TrueType fonts, for as long as I remember. (For example, MessageBoxA accepts char* in the current system encoding, like cp1251 for Cyrillic, and MessageBoxW accepts wchar_t* which is UTF-16.) So, no, it's most definitely not that. It's most likely the fact that people use too much abstraction to achieve simple things.