Bill Moseley wrote Sat, 18 Feb 2006 16:17:20 -0800: > http://www.networksecuritytoolkit.org/nst/tools/css_examples.html > That 1 inch box is a bit larger that 1 1/8 inch on my display. I'm > wondering why. Is it 1" on your display?
That's entirely dependant on the accuracy of the working DPI. You'll find most users will have it display less accurately than yours. > Assuming my display was setup incorrectly, I checked my monitor size > and resolution : > 1280x960 and 19.75" diagonal. > Says I'm 81 DPI. My Xserver logs also say I am 81 DPI: > (**) MGA(0): Display dimensions: (400, 300) mm > (**) MGA(0): DPI set to (81, 81) > And, sure enough, if I draw a 81 x 81 pixel box on the screen it's > exactly 1" square. > If that box is not 1" does that also mean that my 12pt fonts are not > really 12pt? Yes, but they're much much closer to 12pt than average for personal computers. Print out http://mrmazda.no-ip.com/auth/PointsDemo.html and hold it up next to your screen for comparison. Your's is rather accurate because you're using an open source OS desktop environment, which is generally much smarter than doze about implementing a realistic DPI. Study http://mrmazda.no-ip.com/auth/dpi.html and it isn't too hard to figure that few resolution and screen size combinations come close to achieving the artifically assigned 96 DPI that doze uses default, or its 120 alternative common on high resolution laptops. Now whether you actually want your DPI to be accurate is another matter entirely. You may find that object sizing in your desktop environment is less than what you might consider ideal. You can force X to use any particular DPI you please, which will force DPI dependant objects up or down in size to suit your taste. > I'm using relative font sizes on my sites -- letting the browsers > default set the initial font size. Seems like browsers often have > large sizes by default. For those browsers using defaults assigned in pt, they wouldn't actually be "large" if the DPI was accurate instead of an assumed value in excess of reality. 12pt should be the same 12pt you get in print, but typically isn't, because the operating DPI in most cases exceeds reality. Looking back to my chart above, you can see at the XP default 800x600 when used on a 15" nominal/14" real CRT is running at a real 71 DPI, which the assumed 96 is overstating by 35%. This is why experts familiar with the subject frown on the use of pt for CSS sizing of screen media objects. Other browsers use a finer grained default, px. They aren't at all DPI dependant, so whether they're in fact "large" or not depends on display size and screen resolution. > Assuming that many (most?) users don't > change their default font sizes, is there a list of what the > different browsers use for their defaults? This should be about as close as you'll find anywhere: IE most 12pt (16px) IE high resolution (commonly laptops) 12pt (20px) Opera most 12pt (16px) Opera high resolution 12pt (20px) Konqueror 11pt (~15px @ 96 DPI) Tier 1 Gecko (Firefox, Mozilla, Netscape 6+, SeaMonkey) 16px Other Gecko (a few, there are others): Camino might be 14px Epiphany 11pt/~14px K-Meleon 16px Safari (I think same as tier 1 gecko at 16px, possibly 12pt/16px) Note that recent Gecko browsers have a hidden pref browser.display.screen_resolution that may enable you to have it to run at a more accurate DPI than what the desktop uses. You can use about:config to change it and see the effect with each change after restarting Gecko. Note that on pages such as http://mrmazda.no-ip.com/auth/dpi-screen-window.html the reported DPI may be off anywhere from 0 to 12px due to Gecko's broken internal rounding method. -- "Love your neighbor as yourself." Mark 12:31 NIV Team OS/2 ** Reg. Linux User #211409 Felix Miata *** http://mrmazda.no-ip.com/auth/auth ______________________________________________________________________ css-discuss [EMAIL PROTECTED] http://www.css-discuss.org/mailman/listinfo/css-d IE7b2 testing hub -- http://css-discuss.incutio.com/?page=IE7 List wiki/FAQ -- http://css-discuss.incutio.com/ Supported by evolt.org -- http://www.evolt.org/help_support_evolt/
