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

RE: [css3-background] clarify which properties in this module apply to ::first-letter and ::first-line

From: Brian Manthos <brianman@microsoft.com>
Date: Tue, 3 Apr 2012 20:07:17 +0000
To: fantasai <fantasai.lists@inkedblade.net>
CC: "www-style@w3.org" <www-style@w3.org>
Message-ID: <9710FCC2E88860489239BE0308AC5D1712979FE4@TK5EX14MBXC264.redmond.corp.microsoft.com>
That's the kind of deeper evaluation and clarification that I think is warranted.   Thank you.


Next question:

When a customer reports an "IE9 bug" because we enforce the about-to-be-removed conformance requirement and thus their page renders differently in IE9 vs. other browsers, where will the rationale below be captured for reference?  Is there a place where such valuable information is stored?  Or should I rely on the archive for this mailing list as the only official reference.

In addition to aiding customers in understanding, such reference information is also important as we evaluate whether to change future implementations of IE to expand the functionality to leverage the proposed "may" for interoperable and competitive reasons.


Thanks,
-Brian

-----Original Message-----
From: fantasai [mailto:fantasai.lists@inkedblade.net] 
Sent: Tuesday, April 03, 2012 12:09 PM
To: Brian Manthos
Cc: www-style@w3.org
Subject: Re: [css3-background] clarify which properties in this module apply to ::first-letter and ::first-line

On 04/03/2012 10:39 AM, Brian Manthos wrote:
> It's more than that.
>
> The requirement to NOT allow it for box-shadow was explicitly stated.  Has the due diligence been done to research why that was the case?

As I said, we attempted to align box-shadow with the CSS2.1 conformance requirements on borders because, like borders, it delineates the box.
The result was wrong for two reasons:
   a) The reason borders were excluded in 2.1 wasn't because they delineate
      the box (as backgrounds also do this), but because they affect layout.
   b) Our wording forbid 'box-shadow' on ::first-line, whereas 2.1 allows
      borders to apply through the "may apply other properties" clause.

> How many pages on the planet would be broken by changing that?

I'd wager none. 'box-shadow' on ::first-line doesn't seem like an especially common thing to try.

> Seems like an another post-CR arbitrary change to CSS3.

We're supposed to correct errors. Given CSS2.1 allows UAs to apply more than the properties listed, and the CSSWG never explicitly discussed and resolved on this, I consider outright forbidding 'box-shadow' on ::first-line to be an error.

Whether we allow it or not depends on the CSSWG's resolution on how to handle ::first-line properties going forward: on whether we choose to make ::first-line as permissive as possible or as restricted as possible.
And if we intend for Level 4 to require box-shadow to ::first-line, then we certainly should relax the restriction in L3 so that it is not non-conforming.

> There's a larger story here also:  Don't put conformance requirements into the spec that are frivolous.  It's a waste of implementer resources to enforce requirements that you end up casually waving off later.

Nobody's perfect. You don't release software free of bugs; neither do we release specs free of issues. Of course we try, but it's unrealistic to expect we will succeed.

~fantasai
Received on Tuesday, 3 April 2012 20:08:13 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 17:20:52 GMT