W3C home > Mailing lists > Public > www-style@w3.org > January 2011

Re: [css3] support for filters

From: Dean Jackson <dino@apple.com>
Date: Thu, 6 Jan 2011 12:43:28 +1100
Cc: Rik Cabanier <cabanier@adobe.com>, Charles Pritchard <chuck@jumis.com>, ed@opera.com, www-style@w3.org
Message-Id: <1FDCC432-7E65-4EA8-8CC4-3914C93A7032@apple.com>
To: Cameron McCormack <cam@mcc.id.au>

On 06/01/2011, at 11:31 AM, Cameron McCormack wrote:

> Rik Cabanier:
>> Thanks for the info!
>> Having the filter defined externally to the CSS by using 'url(#..)' will make it harder to use. Is there a proposal to make it part of css?
>> Animated filter are very useful ie it's very common to use a blur filter to simulate motion.
> I think it is worth looking into having shorthand filters specified just
> in the property itself, like you have with blur(5) below.  If you wanted
> to support all of what SVG filters currently supports, by writing
> something in a single property, it’s going to get somewhat complicated,
> though.  I don’t know how far we want to take it.

Yeah, it could get ugly defining a filter graph in a property. It would probably end up looking like lisp code - you'll be lost in the parentheses. Also, it's not exactly straightforward to write complex filter chains anyway. In many cases a set of built-in effects will be what developers want.

> An advantage of allowing filter definitions in the filter property, as
> you have done, is that CSS Transitions can define what it means to
> interpolate between two values.



>  If we only have url() values, then it’s
> just going to be a discrete change, and you would need to have the
> transition/animation targetting the <feGaussianBlur> element itself.
> And that would mean that if you want to have multiple elements having an
> animated blur, you would need separate instances of the whole <filter>
> so that they could separately animated their blur radius attributes.
> -- 
> Cameron McCormack ≝ http://mcc.id.au/
Received on Thursday, 6 January 2011 01:44:36 UTC

This archive was generated by hypermail 2.4.0 : Friday, 25 March 2022 10:07:54 UTC