This entire topic is why I DREAD the coming of web fonts – and why 99% of the websites I develop stick with this:
font:normal 85%/140% arial,helvetica,sans-serif;
In the body CSS, and doesn’t get TOUCHED for anything other than headings or image interactions.
You’ll constantly hear mimicry of the lame excuses from the “I learned print and they called it web” PSD jockeys who fail to grasp that the web is NOT print. All their fancy serifs and fancy typography does on the web is get in the way of delivering the content to the user – defeating the point of even having a website in the first bloody place. I don’t care how pretty your goofy font is, if it prevents anyone from reading the content
The same goes for colors when the art *** up and decides on letting their inner child go wild – now, the WCAG’s formula for checking contrast is a bit complex… but that’s why I personally go with the old IBM/MS/APPLE co-developed contrast norms they put together for the old Win3, OS/2, and MacOS norms way way way back in the day – 50% luminance minimum, 75% luminance preferred. The formula for luminance on that measurement being:
Which is hard-coded into the VGA for converting color to a monochrome display as well as the ATI chips used in most early color Macs, and can be found in several old computer books, most notably “The Programmers Guide to the EGA/VGA” by Richard E. Ferraro, “The Windows User Interface Guidelines” included in PDF form with the old Visual Basic and VC++ for Win 3.x, and there was an Apple Book for Object Pascal that said the exact same thing… “If the user can’t read your text, what good is it?”
If you take the foreground color, convert it to L, take the background color, convert it to L, and they are more than 50% apart they’re probably illegible to most people – and if they are 50-75% apart some people may have difficulty depending on the colors. (light on dark is more forgiving than dark on light in that regard).
… and the best part is that formula also miraculously prevents color-blindness issues too!
But it’s like font-size… as a large fonts/120dpi user I find anything smaller than 12px illegible, and even 14px is on the edge – which is why when FORCED to use px by an image interaction I go with 14px or larger, but I ALWAYS keep my content at 85% for sans-serif and 100% when/if serif is appropriate (which is to say almost never).
… and yeah, I’m not a fan of tag clouds either – what’s wrong with just listing them in order of popularity? OH wait, that would make SENSE.
Oh, and @ScallioXTX – You know, I love it when people link to articles talking about print, that are themselves illegible on screen THANKS to the use of serif fonts. Screen is NOT print – that whole 6.2% to 30% the resolution thing being a real kick between the legs on the use of serifs.