Re: vendor prefixes: co-cascading

The integration features Microsoft introduced for Windows 7 site-pinning 
will never become a standard because it's a platform-specific feature. The 
same goes for the Apple-pinning icon. Those things are very useful, but they 
simply don't make sense in another context that the one they were made for. 
The same story goes again for Microsoft Word, which use "special" CSS 
properities to store things that are not representable in CSS today but they 
want to retreive back when the HTML is opened in Word later. You can find 
similar properties in "Prince", an HTML bookprinter which need ways to 
represent headers and footers in a way the CSS spec doesn't allow at this 
time. All those things are platform-specific and should be prefixed (and 
they are, which is nice).

Experimental prefixes are useful in some contexts. Trying to kill them 
completely is a waste of time, they'll never disappear.

I'd prefer the CSS WG to concentrate on fixing the most important concern: 
stopping them to propagate in the "big whole web" for features *that are 
being standardized*. Features that have been rejected for standardisation or 
need to be implemented for platform-specific reasons should continue to use 

-----Message d'origine----- 
From: Henri Sivonen
Sent: Friday, November 18, 2011 11:37 AM
Subject: Re: vendor prefixes: co-cascading

On Fri, Nov 18, 2011 at 12:21 PM, François REMY
<> wrote:
> If you both don't want prefixes in non-RTM browsers and advocate that
> experimental features should be kept in non-RTM browsers, then we are 
> going
> for a model where nobody prefixes anything (exepept platform-specific
> things). Is it what we really want?

Why not?

What would be a legitimate platform-specific thing to introduce?
Surely the a priori assumption of anyone who is introducing a feature
and believing the feature to be Good for the Web should be that the
feature will be eventually implemented by everyone. Maybe others will
disagree and outright refuse to implement the feature, but still, the
assumption one should start with when introducing a new feature is
that it's so good that everyone should want it. If the introducer of
the feature doesn't believe the feature to be so good, they shouldn't
be introducing the feature.

Henri Sivonen 

Received on Friday, 18 November 2011 12:32:50 UTC