- From: David Perrell <davidp@earthlink.net>
- Date: Thu, 24 Apr 1997 13:53:54 -0700
- To: "Chris Lilley" <Chris.Lilley@sophia.inria.fr>, <wmperry@aventail.com>
- Cc: "Paul Prescod" <papresco@calum.csclub.uwaterloo.ca>, "Todd Fahrner" <fahrner@pobox.com>, <www-style@w3.org>, <Jukka.Korpela@hut.fi>
Chris Lilley wrote: > That is not the meaning of relative font sizing in CSS1, as you are > well aware. Sor-reee. My point is that "absolute" measurements aren't. By spec'ing type in point sizes commonly used for print, the displayed result is likely to be translated into sizes that are comfortable for the viewer. The result is more likely to be legible than measurements based on pixels, but unlike print it won't match the measure. > I would direct you to the definition of a pixel in CSS, as well. I don't believe UAs are going to be considering pixels so philosophically any time soon. > So to get back to the original question - current usage tends to go > for absolute rather than relative font sizing, I agree. This is partly > an education issue and partly a case of people working around incomplete > implementations. Also the widespread use of old X servers that use > bitmapped fonts rather than scalable fonts may be a factor - people > pick sizes they know they have good bitmaps for. When you've been spec'ing type in points for eons, you know what numbers produce the effect you're looking for. But fossilization isn't the only impediment to use of relative values. Design and layout are about relationships. Consider an inline graphic that must be sized relative to the surrounding text. How is this done if the parent element's font is "120%" or "larger"? I'll save you the trouble of answering: with em and ex. As you know, these units aren't supported in early implementations, and their usefulness is sometimes not apparent to those for whom typography is irrelevant. But for anyone designing truly cascading stylesheets, em's essential. David Perrell
Received on Thursday, 24 April 1997 16:58:54 UTC