Re: Info from HTML5 editors: merging WHATWG patches

On Sun, Sep 9, 2012 at 9:32 PM, Silvia Pfeiffer
<silviapfeiffer1@gmail.com>wrote:

> On Sun, Sep 9, 2012 at 11:01 PM, Glenn Adams <glenn@skynav.com> wrote:
> > On Sun, Sep 9, 2012 at 4:49 PM, Silvia Pfeiffer <
> silviapfeiffer1@gmail.com>
> > wrote:
> >>
> >> (B1) branches with fixes for typos or that resolve bugs in our bug
> >> tracker (i.e. we likely all appreciate that these should be applied).
> >> (B2) branches with features that are either new, or for which I don't
> >> know if we should merge them.
> >> (B3) branches/patches with features that we decided to hold back from
> >> HTML5 (also listed in [5]).
> >
> >
> > I believe that, at this point in the process, any change that either (1)
> > introduces a new feature or (2) would make a substantive technical change
> > for which there is no related LC bug where that change is an agreed
> > resolution should have explicit approval (i.e., resolution to that
> effect)
> > by the WG.
>
> If the world were so easily black an white, it would be simple.
>
> > For me, a substantive change in this regard is any change that may affect
> > conformance or constitutes a change at a syntactic level, e.g., a WebIDL
> > change or a change to the set of defined elements, attributes or
> attribute
> > values.
>
> I am assuming where such a WebIDL change is made because the spec is
> different from browser implementations, it is a minor spec bug that is
> being fixed and not worth a new feature branch. It will still be in
> one of the branches classified as B1.


Some WebIDL changes are more substantial than others. I would grant that
adding a [TreatNullAs=EmptyString] is not necessarily worth bringing to the
WG. On the other hand, adding a new interface member, changing its
signature or type, etc., is worth bringing to the WG's attention (in some
manner).

Simply assuming that there is agreement among UA vendors for a given change
is not sufficient unless there is some documentation trail pointing to that
agreement. Further, the WG's members aren't simply UA vendors. Author tool
vendors and content authors also have a significant stake in the WG's
process and its results.

I'm merely suggesting that the editors exercise caution (erring on the
conservative side) when making clearly substantive changes if it falls into
the categories I described above unless the WG has had an opportunity to
weigh in on the matter. On the other hand, I wouldn't want to see every
change that is in a clearly gray area to take up the WG's time or process.

I'm willing to agree to the editors making substantive changes they think
are non-controversial, as long as there is a list of such changes available
during the next LC or CR in a "Change History" section.