W3C home > Mailing lists > Public > www-style@w3.org > May 2004

Re: Colour gradient backgrounds.

From: Ernest Cline <ernestcline@mindspring.com>
Date: Sat, 15 May 2004 23:45:28 -0400
Message-ID: <410-22004501634528312@mindspring.com>
To: "Lachlan Hunt" <lachlan.hunt@iinet.net.au>, "W3C Style List" <www-style@w3.org>


> [Original Message]
> From: Lachlan Hunt <lachlan.hunt@iinet.net.au>
>
> LĂ©onie Watson wrote:
> >     SVG does present a workable alternative, but it still has minimal
> > support, which I think someone else already mentioned. SVG is
> > likely to take some time before it reaches critical mass in terms
> > of support and implementation, something css is already well in
> > advance of.
>
>    Lack of support for an existing standard, such as SVG, which can 
> already produce the desired effects, is no reason start filling in the 
> gaps in CSS, using new properties which will also take quite some
> time to get widely implemented.

Just to make it clear, the reason I have for wanting CSS gradients
is not as a stopgap until SVG is widely implemented, but as
something that would be useful even in environments where SVG
will never be implemented.  The only consideration my proposal
made with regard to SVG, was to make certain that its adoption
would not cause problems for SVG.  SVG largely ignores the box
model, my gradient proposal is dependent upon it.

<snipped section by Lachlan about paralleling every background
 property with a corresponsing foreground property.>

It's not a bad idea in the abstract, but it suffers from one major
problem: property inheritance.  This problem isn't insurmountable,
as I had to face it for the foreground gradient property in my
proposal,  It's why the 'gradient' property  used the "auto" value
as a default value that caused a virtual inheritance of a gradient
from the closest ancestor that  had one defined.  One could
change 'color' to do something of the same sort, but it would be
awkward compared to actual inheritance and I'm not certain that
it would be 100% compatible in effect with how 'color' works now.
(I think it would, but I'm not certain.) The most likely area for
incompatibilities from a change in 'color' from real to virtual
inheritance would be in scripting.  It is also possible that to
achieve the desired effects of 'foreground-image', etc. one could
leave 'color' alone.  If that is possible, it would be desirable.

There is also a minor problem, in that 'color' would have multiple
shortcut properties ('font' and 'foreground') that would reference it,
a difficulty that so far has only affected border properties.

In short, both foreground gradients and foreground images are
something that could be done in CSS, but they are more
complicated than the background versions to implement
properly and achieving 100% similarity between how
backgrounds and foregrounds work is not achievable.
Received on Saturday, 15 May 2004 23:46:20 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:54:30 GMT