Re: [css3-background] Restore box-decoration-break

On Tue, Mar 4, 2014 at 8:09 PM, Sylvain Galineau <galineau@adobe.com> wrote:

>
> On Mar 4, 2014, at 5:08 PM, Glenn Adams <glenn@skynav.com> wrote:
>
> >
> > On Tue, Mar 4, 2014 at 2:26 PM, Sylvain Galineau <galineau@adobe.com>
> wrote:
> >
> > On Mar 4, 2014, at 11:45 AM, Glenn Adams <glenn@skynav.com> wrote:
> >
> > >
> > > On Tue, Mar 4, 2014 at 12:31 PM, Simon Sapin <simon.sapin@exyr.org>
> wrote:
> > > On 04/03/2014 19:06, Glenn Adams wrote:
> > > Cox would like to see box-decoration-break restored to this
> > > specification rather than postponed to Level 4.
> > >
> > > Hi,
> > >
> > > The property is not postponed to Backgrounds and Borders Level 4, it
> has been moved to the Fragmentation module:
> > >
> > > http://dev.w3.org/csswg/css-break/#break-decoration
> > >
> > > ok, thanks for correcting that; in any case, Cox is concerned that
> moving it into another spec will delay it being published in a final (REC)
> form, as compared to leaving it in B&B3
> > >
> > >
> >
> > Given that REC requires 2+ implementations passing test cases for the
> feature, this concern suggests we have 1) sufficient test case coverage of
> box-decoration-break to establish that 2) there is a good interop. Do we
> have 1) and the data for 2) ? Or is Cox maybe offering to share the
> testcases in 1) that confirm 2) ?
> >
> > You're putting the cart before the horse.
>
> This statement is neither helpful nor accurate. Exiting CR requires a test
> suite and 2+ implementations passing each test. If you have testcases that
> lead you to believe box-decoration-break is as close - or closer - to
> interop than the other widely deployed and stable features in B&B then your
> concern is legitimate. If not then your concern is not really actionable.
>
> > The CR period is also known as "Call for Implementations". The SOTD
> section notes that "A test suite and implementation report for CSS
> Backgrounds and Borders Level 3 will be developed during the Candidate
> Recommendation phase, which will last a minimum of three months." Let's get
> into CR before making conclusions about what tests suites are available and
> what is implemented.
>
> As I am not the one expressing concerns about a specific feature missing
> REC it seems rather odd to imply I am prematurely jumping to conclusions.
> Second, B&B is not some brand new spec we started working on last year. It
> is quite mature and defines a number of features web content already
> depends on today. Thus adding features to B&B CR that have shakier or even
> no UA support implies delaying REC for the rest.
>
> As such I believe it is up to Cox to back up its concern with the evidence
> that led it.
>
> >
> > I think it’s great if box-decoration-break is interoperable; but if it’s
> not there yet, holding back REC for a number of B&B features with better
> cross-UA support may not be a better trade-off.
> >
> > Nothing is holding back REC except our rather slow process of getting
> specs moved to CR.
>
> You either misunderstand our ‘slow process’ or are missing my point. Maybe
> both. See the previous comments.
>
>
We don't have a B&B CR yet, so what are you talking about? The document is
in LC, it has not yet started CR.

Let me explain how W3C process works since you seem unfamiliar with it:

(1) you publish a LC
(2) you resolve LC comments
(3) you publish a CR and call for implementations
(4) you publish a test suite and assess implementation reports
(5) you identify features at risk
(6) you remove features at risk if needed based on test suite support and
interop reports
(7) you publish a PR
(8) you publish a REC

At present were are (1) [1] - that there was a previous CR for a now
out-of-date version in 2009 [2] is not relevant. We are not at (6). Now,
the phrase "put the cart before the horse", which you also seem unfamiliar
with, is "an analogy for doing things in the wrong order" [3].

Since you are making an argument about the presence of tests and
interoperability (4) in order to justify removing this feature as a feature
at risk (5)(6), but we are now in step (1), you are clearly doing things in
the wrong order.

Is that clear, or shall I resort to first principles?

[1] http://www.w3.org/TR/css3-background/
[2] http://www.w3.org/TR/2009/CR-css3-background-20091217/
[3] http://en.wikipedia.org/wiki/Cart_before_the_horse