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

RE: Proposition to change the prefixing policy

From: Sylvain Galineau <sylvaing@microsoft.com>
Date: Sat, 5 May 2012 23:04:35 +0000
To: Daniel Glazman <daniel.glazman@disruptive-innovations.com>
CC: "www-style@w3.org" <www-style@w3.org>
Message-ID: <3C4041FF83E1E04A986B6DC50F0178290A349160@TK5EX14MBXC261.redmond.corp.microsoft.com>

[Daniel Glazman:]
> 
> Le 05/05/12 16:03, Sylvain Galineau a écrit :
> 
> > *: Yes, I am thinking of gradients. Yes, I agreed to most of these
> > changes at the time. Yes, I now believe I was mistaken. Yes, we
> > shouldn't extrapolate from just one recent case. Still, the process
> > explicitly allowed and enabled these changes to happen. I see no
> > reason to believe this can't happen again beyond the fact that some of
> us acquired a strong distaste for such decisions from that experience.
> 
> Then you probably think of CSS Transforms. It has had skew() for ages,
> it's used in the wild, and then the last version of the spec got rid of it.
> After 3 years. Hum ?
> 
No, I do not. A query against an internal database of ~1,000,000 resources 
from ~100,000 web sites selected by traffic returns 27 stylesheets using this 
function on 25 sites. In terms of the number of stylesheets stored in the 
system, this works to about 0.000078% of the stylesheets we've crawled in this
particular snapshot of the web.

Even if we include other content types to account for script elements and 
JavaScript resources (which may produce false hits as a function called
skew will be a hit) we only get to about 0.1% of all resources. Granted, 0.1%
of the web is big but, again, keep in mind that a lot of these hits are false
positives.

By comparison, the same query on rotate() returns orders of magnitude more
and impacts nearly 13% of the domains stored. So which one can we remove or
change without causing headaches?

While this is only a, um, skewed sample of the entire web I'm thus comfortable 
with claims that skew() achieved too little traction over 3 years to matter. 

Last, if you're interested, linear-gradient is close to as prevalent as rotate()
in this same web snapshot. Sadly, this database is not public and I do not expect
it will be soon. But I hope this can at least clarify why I'm comfortable with 
certain changes and deeply regret others.

Received on Saturday, 5 May 2012 23:05:13 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 17:20:53 GMT