- From: David Perrell <davidp@earthlink.net>
- Date: Tue, 20 Nov 2096 13:29:27 -0800
- To: "Stu Harris" <sirrah@baluga.maximumaccess.com>, "Stefan Olson" <stefan@olsonsoft.co.nz>
- Cc: <www-style@w3.org>
Stu Harris wrote: > ... Pixel measurements are the only reasonable way of thinking > about relationships on a Web page. Quite apart from the fact that the > vast majority of the world measures things in centimetres. I disagree. As far as legibility goes, pixel measurements mean less than inches. On the same size display, 10-pixel text legible at 640 x 480 can be illegible at 1600 x 1200. Mapping of display pixels/inch should be settable in the opsys (ideally for both x and y). Given a 9"H x 12"W display at 640 x 480, for example, 53 pixels per inch is 'actual size'. At 1600 x 1200, 133 pixels/inch is actual size. At 1280 x 1024 it would be 107 pixels/inch horizontal, 114 pixels/inch vertical. Once display mapping is set, it doesn't matter whether you spec in points, inches, or centimeters, the measurements will be accurate. In Windows, pixels/in is a function of the display driver. Some drivers allow some control. I have a 16"-wide (21" diag) 1280 x 1024 display, with display pixels/inch set to 112. This is larger than actual size because the monitor is farther away than I would normally read printed material. So when 12pt type is spec'd using CSS1, the type appears in MSIE 3.01 as it would on paper at a comfortable reading distance. All this is moot as far as CSS is concerned, since measurements can be spec'd in CSS1 many ways including pixels. But if you want legible text on the widest variety of displays, it is better to spec in points than pixels. David Perrell
Received on Wednesday, 20 November 1996 16:35:17 UTC