[nxn]

So to get this straight, the author created a button image in photoshop, then created the same style using css3, took a picture of both, zoomed in, and compared the pixelated blockyness? ... While completely disregarding the fact that actually zooming in on the css3 version in the browser would avoid any pixelation and blockyness in the first place?

[ryanf]

He's zooming in so you can see the pixels. He obviously isn't talking about putting giant, stretched-out buttons on an actual site.

[nxn]

Which is not what I'm talking about either. I'm just saying it's a bit ironic to need to zoom in on a button to even distinguish any quality differences and then giving the quality award to the method that isn't suitable for scaling.

[skalpelis]

The zoomed image was meant to illustrate why the photoshop button was rendered better at default size - it has nothing to do with browser zooming.

[nxn]

I understand that, in fact I mentioned that the aspect of zooming was completely ignored in my first post, and I found it ironic that it was.

But anyway, lets forget about that fact for now. If you need to zoom in 400-800x on some pixels to even be able to spot the differences, then in practical terms there are no differences. You can't expect your typical visitor to sit in front of the monitor with a magnifying glass sweating at your pixel perfect buttons. Honestly, no one is going to give it much more than a glance, so attention to detail you can hardly spot with the naked eye is almost completely pointless.

Now to come back to my main point: what isn't pointless is maintaining quality when zooming in with the browser. This is something I often do myself when I'm too lazy to put my contacts in to just casually browse the web. Just about every website looks like a mess of blurry/pixelated crap when you do that. Had he just went with the css3 approach, the quality at non standard zoom levels would far outweigh the minuscule pixel details at normal zoom levels between the css3 and photoshop versions.

Anyway, that's all I really wanted to say. Had the point been that CSS3 styling was still iffy with older browsers being around and the photoshop button was superior for that reason, I wouldn't have said anything. But to dismiss it based on some differences in pixelation you wouldn't even spot normally, while ignoring that if you had actually zoomed in the browser there would be no pixelation at all with the "inferior" css3 version, was just too much.