Barney Carroll wrote: > > …FF also does subpixel positioning while Safari snaps everything to > > pixel positions, as Philippe noted. > > Briefly, while we're on the subject, I might add that this difference of > policy of maths-to-the-pixel applies to a great number of things (not > just letter-spacing) — in one of the less dynamic battles in the war for > good typography on the web, it might be noted that many (of the hundreds > of otherwise quite beautiful) badly-hinted fonts can become quite > passable in Firefox when given a text-shadow in the same colour as the > font, with offsets at 0 and a range/size of .6 pixels, thereby giving an > anti-aliasing compensation for otherwise blocky presentation. The same is > sadly not true of Webkit — ditto for cumulative fractionally defined (%, > em) metrics (use case: columns) and far more. > > Ironic considering Webkit has for some time been the de facto standard > browser on variable pixel-density devices, where the issue becomes > incredibly pertinent.
Right, and this has spilled over into the design of CSS3. I was pushing for kerning to be on by default but the WebKit folks felt this would be too slow, based on their testing. So the current CSS3 Fonts spec [1] includes this new property: font-kerning: auto | normal | none; The value normal means kerning is on, none means off and auto means "browser decides" (the default will be auto). For FF I think we will concentrate on making the kerning path fast rather than making the non-kerning path fast. Just a different philosophy/focus I think. Regards, John Daggett Mozilla Japan [1] http://dev.w3.org/csswg/css3-fonts/#font-kerning-prop ______________________________________________________________________ css-discuss [css-d@lists.css-discuss.org] http://www.css-discuss.org/mailman/listinfo/css-d List wiki/FAQ -- http://css-discuss.incutio.com/ List policies -- http://css-discuss.org/policies.html Supported by evolt.org -- http://www.evolt.org/help_support_evolt/
