W3C home > Mailing lists > Public > www-style@w3.org > February 2012

Re: [css3-background]

From: Marat Tanalin | tanalin.com <mtanalin@yandex.ru>
Date: Wed, 29 Feb 2012 03:17:45 +0400
To: Tab Atkins Jr. <jackalmage@gmail.com>
Cc: Lev Solntsev <greli@mail.ru>,"www-style@w3.org" <www-style@w3.org>
Message-Id: <158291330471065@web155.yandex.ru>
A possible option is to allow rgba() as direct background-image property value -- without redundant image() function-wrapper.


28.02.2012, 19:57, "Tab Atkins Jr." <jackalmage@gmail.com>:
> On Tue, Feb 28, 2012 at 2:11 AM, Lev Solntsev <greli@mail.ru> wrote:
>
>>  Of course I can use gradient, Image() or whatever. But isn't it a hack and
>>  workaround while there is a more obvious way for authors, is it?
>>  Also I have concerns about performance issues in browsers and
>>  those concerns contradict to accessibility—what web was created for.
>>
>>  http://www.w3.org/Style/CSS/ says that “Cascading Style Sheets (CSS)
>>  is a simple mechanism for adding style (e.g., fonts, colors, spacing)
>>  to Web documents.”
>>
>>  Using any of these workarounds make things more complex than it could be.
>
> Lea gave the main reason why it would be difficult to define - if you
> have linked up property-lists, like in the background-* properties (or
> the transition-*, or the animation-*, etc.), you typically link up the
> corresponding entries in order to form a full set of properties.
> background-color doesn't work like this - if you specify a single
> background-color, you want it to match up with the *last* background
> layer, not the first.  This is different from how property lists work
> everywhere else in the language.  It's also unclear what would happen
> if you specified multiple background-colors - would you start matching
> them from the front instead, or continue matching them against the
> end, so the first color is placed in the second-to-last background
> layer, etc.?
>
> We could have, several years ago when multiple background layers were
> first introduced, made background-color consistent, but that would
> mean that the vastly more common case (just wanting a single color at
> the end) would require a long list like "background-color:
> transparent, transparent, transparent, transparent, blue;".
>
> So, this is an optimization problem.  The simplest solution of just
> making background-color work just like the other properties (which we
> can't do anymore anyway) would be very inconvenient in the vast
> majority of cases.  A more complicated solution making
> background-color "do the right thing" when you give a single color
> would be inconsistent with the rest of the language, and thus hard to
> learn.  Instead, we hacked around the issue and created ways of
> producing solid-color images, which is a simple way to solve your
> reasonably rare case.  It's not perfect, but it's the least bad of the
> solutions (in our judgement).
>
> ~TJ
Received on Tuesday, 28 February 2012 23:18:17 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 17:20:51 GMT