3 ms·
I'm skeptical that most websites would see a performance improvement in switching from images to symbol fonts. Github is doing it close to optimally: - custom
by cbr 14y ago
I'm skeptical that most websites would see a performance improvement in switching from images to symbol fonts. Github is doing it close to optimally:
- custom webfont that has exactly what they need (26k)
- loaded in js to minimize downloads (only one of woff and eot)
- served with very long cache lifetime
- served via cdn
But even then they have the problem that all text display has to wait for the font to download.
Most sites are going to skip some of these, so I expect the change to symbol fonts to hurt most websites that try it.
If you do try it, definitely A/B test it to make sure it really is speeding up pageloads for your users.
- crazygringo 14y agoFor me, the #1 concern is mobile zooming and retina displays. No more blurry pixellated edges that suddenly make your site look like total crap. Instead, it's as sharp as a printed page! For me, improving page loading speeds are a secondary concern -- and as long as symbol fonts don't make the page noticeably slower to load, all is good.
- cbr 14y agoI agree: the vector sharpness of symbol fonts is their strongest selling point. I'm just worried that the article leading with "improving performance is a constant process" is going to make people think symbol fonts will speed up their pages.