- From: fantasai <fantasai.lists@inkedblade.net>
- Date: Thu, 10 Sep 2015 19:00:38 -0400
- To: Greg Whitworth <gwhit@microsoft.com>, Florian Rivoal <florian@rivoal.net>
- Cc: "www-style@w3.org" <www-style@w3.org>, w3c-css-wg <w3c-css-wg@w3.org>
On 09/10/2015 05:17 PM, Greg Whitworth wrote: > > 2. Please change the following: > > # 3.3.3 Proprietary and Non-standardized Features > # Example 2 > # For example, Win8 Metro apps... > > To... > > "Microsoft, Apple, and Mozilla, among others, have utilized > extensions to standardized APIs implemented by their respective > UAs that do so without allowing web content to access these > features. Because of this, it alleviates the opportunity for > such content to become dependent on their proprietary extensions." I'm not opposed to changing the example text, but I think the wording you propose is very abstract, and it helps in examples to be specific and concrete. > 3. Please change the following: > > # 3.3.2 Market Pressure and De Facto Standards > # If at least three... > > If I'm not mistaken our CR Exit Criteria is only 2 and # we also believe that it would only take two browsers > (depending on the share maybe even one) implementing > something that becomes heavily utilized by authors to > become a de facto standard which should give other UAs > the blessing of the CSSWG to implement for web compat. > Due to this, we suggest changing this to: > > If at least two... It's three here because we're dealing with a case where a) there isn't a content dependency, because *nobody* has shipped an implementation for broad use yet--they're all behind flags or whatever b) the spec isn't stable (i.e. not in CR, or otherwise agreed to be done enough to reliably implement against) The case you're talking about where there is one (or more, but even just one) implementation that is heavily used falls under the parenthetical: # (or if a browser has broken the previous rule and shipped # for broad use an <a>unstable</a> or otherwise non-standard # feature in a production release) and in this case the number of implementations required to trigger the clause is just one (though preferably more). Is that making sense, or am I failing to write clearly? :) ~fantasai
Received on Thursday, 10 September 2015 23:01:13 UTC