- From: Simon Sapin <simon.sapin@exyr.org>
- Date: Mon, 25 Mar 2013 13:20:09 +0100
- To: www-style@w3.org
Le 25/03/2013 05:58, L. David Baron a écrit : > http://dev.w3.org/csswg/css3-text/#tab-size defines a 'tab-size' > property taking an integer to be multiplied by the advance width of > the space (U+0020) character. > > However,http://dev.w3.org/csswg/css3-values/#font-relative-lengths > defines the 'ch' unit as using the advance width of the '0' (U+0030) > character, which I believe was the result of implementation > experience that '0' produced good results for sizing in proportional > fonts, and made no difference in monospace fonts. (I think that > implementation experience related specifically to sizing of text > inputs.) > > Should 'tab-size' instead use the '0' width rather than the ' ' > width? I think I would prefer such a change. I like that this simplifies implementation (<integer> values are either invalid or converted to <length> with the ch unit) but I wonder about the impact on existing content. In my browser/system’s default font (I’m not sure which it is), 1ch is about twice the width of a space. Assuming the initial value remains 8(ch), this change makes a big rendering difference for preserved tabs in non-monospace fonts. Do we have data on how (un)common these are? Test case: data:text/html,<a>0000</a><p><a> </a><style>a{border:1px solid;white-space:pre -- Simon Sapin
Received on Monday, 25 March 2013 12:20:36 UTC