W3C home > Mailing lists > Public > www-style@w3.org > June 2013

[css-text-decor] line-decoration averaging or not

From: fantasai <fantasai.lists@inkedblade.net>
Date: Tue, 25 Jun 2013 12:38:39 -0700
Message-ID: <51C9F1BF.3010302@inkedblade.net>
To: "www-style@w3.org" <www-style@w3.org>
We discussed at the F2F requiring the decorating element to set the
position and thickness of the line, ignoring descendant content.
This is acceptable for underlines on alphabetic-aligned text,
however it doesn't work for centered text, which is the most common
alignment in vertical writing modes.

       []
       []
       []
      +--+
      |  |
      +--+
      |  |
      +--+
      |  |
      +--+

An underline positioned to the smaller text will cut through the
larger text. (We have a similar problem with overlines, however
they are less common.)

David argues that allowing descendants to affect the underline
position means underlines may be slightly inconsistent across
similar elements. However, I think it's a bigger problem if
they sometimes cross through text than if they sometimes aren't
pixel-perfect.

If we're unsure about the behavior we agreed to spec last fall,
then I suggest that we mark the positioning and thickness
undefined, as it is in CSS2.1, take this level to CR, and figure
out exactly what rules we want in L4. I don't think it is a good
for an underline to strike through the text of a descendant whose
only crime is having a larger font-size, and I do not want our
specs to require it.

Proposed text for undefining:

   "The exact position and thickness of line decorations is
    UA-defined in this level. However, the UA must use a single
    thickness and position on each line for the decorations
    deriving from a single decorating box."

Unlike the CSS2.1 text, this doesn't even give any suggestion
of how to resolve the positioning.

~fantasai
Received on Tuesday, 25 June 2013 19:39:14 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 25 June 2013 19:39:15 UTC