3 ms·
My implication was to do something standards compliant and browser-neutral like 1px wide PNGs and (for the super fancy) JavaScript. Avoid non-standard CSS hacks
by metamemetics 16y ago
My implication was to do something standards compliant and browser-neutral like 1px wide PNGs and (for the super fancy) JavaScript. Avoid non-standard CSS hacks entirely. It has a higher chance of breaking and not-working because, as you said, CSS3 is not finalized.
I wouldn't care if it was supposed to be some sort of academic tech-demo, but in the comments the author says:
>On Posterous they are going to be used in many places
edit: moral: JavaScript enabled\disabled is 1 axis of degradation. Non-standard CSS3 hacks have 3 axises: Browser type, browser version, state of CSS3 spec. It's a statistically worse bet.
- coderdude 16y agoDo you actually develop sites with this kind of thinking? I'm shocked that you'd rather use a 1px png (adding an additional http request) and JavaScript to achieve an effect that CSS can handle all on its own. We use browser extensions because specs like CSS3 aren't suddenly turned on one day like a light switch. You can make your layout look nicer on cutting-edge browsers and by the time the old browsers are no longer around and the spec is in widespread use you can remove the extensions. It's trivial to remove CSS from a stylesheet. It's usually a pain in the ass to re-implement CSS functionality from 1px pngs and JavaScript hacks once all the browsers have caught up.
- metamemetics 16y ago>CSS can handle all on its own CSS can't handle it all on its own is the point. And yes I'd rather incur the extra bandwidth of a 1px wide crushed PNG in order to have simpler code and have my site support all-browsers, including ones whose team made good CSS a priority before other browsers did. Rewarding the large monoliths by coding specifically for them simply because of their popularity while punishing teams that actually prioritized CSS is what got us into this mess in the first place. Would you encourage this behavior pre-CSS4? pre-CSS5?