W3C home > Mailing lists > Public > www-style@w3.org > March 2001

Re: opacity in css

From: Chris Lilley <chris@w3.org>
Date: Wed, 28 Mar 2001 18:53:50 +0200
Message-ID: <3AC2171E.121F0A19@w3.org>
To: James Craig <james_craig@Powered.com>
CC: www-style@w3.org, "'patrick@patrickcurry.com'" <patrick@patrickcurry.com>, Natanya Pitts <natanya_pitts@Powered.com>

James Craig wrote:

> some of the ie 'filters' that were left out of the css spec 

There was never any adequate documentation submitted on how these filters
work, which is the first step to standardisation. They were 'left out' of
CSS2 because no technically sound proposal was made to include them.

In the mean time, SVG has developed a filter mechanism which is applied via
CSS properties, is fully defined in an open W3C specification, and has open
source implementations. I suggest you check them out. Nothing prevents
these filters being added to a non-SVG implementation or being used in a
mixed SVG/block-formatted-text implementation.

> seem pretty
> useful (specifically 'opacity')... 

Opacity is indeed useful. Only doing opacity - in particular, suporting the
opacity in existing image formats - via a filter is kind of an ugly hack,
though. An opacity property (or properties, see fill-opacity and

> today i found out that ns6 supports
> opacity but in a different way.

The -moz-opacity property (note the prefix to indicate it is a non-standard
extension) is much more similar to the opacity property being developed for

> IE:
>  filter:alpha(opacity=50);
> NS6:
>  -moz-opacity:0.5;
> of course, neither works in Opera.

You would expect Win/IE-specific or moz-specific extensions to work in

> now at first i thought it was stupid to stick the mozilla name in a css tag,
> but the more i thought about it, the more i liked the idea of them stating,
> "we know it's not in the spec, but here's how we (-moz-/mozilla) would have
> done it."

Yes, exactly.

> it also seems better supported in ns. in ie, to apply a transparency to an
> object (other than an image) the object has to be absolutely positioned.
> regardless of the details, my discussion topic for the group is: what do you
> think about this functionality branching again? ns4 vs ie4 caused me many
> headaches and i see it going the same way again... it would be nice to avoid
> the headaches, any thoughts?

The solution is to have a W3C specified way to do it. In the absence os a
specification, vendor-specific solutions will naturally branch.

Received on Wednesday, 28 March 2001 11:53:54 UTC

This archive was generated by hypermail 2.3.1 : Monday, 2 May 2016 14:26:57 UTC