Re: Differences between the W3C and WHATWG specifications

Hi, David-

Thanks for the considered reply.  This does raise an interesting point 
of tension between W3C and WHATWG, which I'd like to explore, but leaves 
my primary concern unanswered.

L. David Baron wrote (on 6/22/10 12:40 AM):
> On Sunday 2010-06-20 01:11 +0100, Doug Schepers wrote:
>>  So, the aim of the charter to "actively pursue convergence with WHATWG,
>>  encouraging open participation" has clearly been met from the W3C HTML
>>  WG side.  Speaking for myself (not my employer), I would like to see a
>>  clear message from WHATWG that it intends to act with similar good faith
>>  to present a straightforward collaboration with W3C so that the users,
>>  content authors, implementors, and specifiers know what constitutes the
>>  HTML5 spec.
>
> While I think it's true that the HTML Working Group within the W3C
> has tried to collaborate with WHATWG, I'm not so sure that it's true
> of the W3C as a whole.

Again, I think it's useful to look at the recent history of the W3C to 
gauge how well it adapts to meet the needs of the Web community (as well 
as the active W3C participants and the W3C member organizations who fund 
it), especially in terms of how open and collaborative it has been.

Even before the WHATWG formed, W3C had taken a difficult and firm stand 
for the open Web with its Royalty-Free Patent Policy, rather than the 
more closed RAND option; we lost members (and revenue) over that, but it 
was the right thing to do.

In the years since the WHATWG formed, to match the zeitgeist and 
comments by some in the WHATWG, W3C has become much more open and 
accountable, as well.  To list a few points:

* We are much more liberal with our Invited Expert policy (even setting 
aside the unusually large HTML WG), even in the face of our budget cuts; 
we do this because we want to maintain an open forum for participation, 
and we are trying to make it even easier for people to get involved 
going forward.

* In 2004, there were few if any working groups chartered as public 
groups; we began to encourage groups to consider rechartering to work in 
public, and now the clear majority of groups (and not just those that 
the browser vendors participate in) are chartered public, with most of 
the member-only groups effectively doing much of their technical work in 
public.

* It's now common for groups to have even their editor's drafts publicly 
accessible, so people can keep track of day-to-day changes and comment 
on them.

* Several groups have a working policy that no final decisions are made 
at face-to-face meetings, or even on telcons, to be more inclusive of 
those who cannot or choose not to participate outside of email and polls.

* We are more relaxed in adopting new work, and even proactively seek 
out work we think would help the Web community (like the new Audio XG, 
or the Social Web XG).

* We have increased our focus on testing, with more complete test suites 
in general, and strive to have at least one of the user agents on the 
implementation report be publicly available.

* We put more emphasis on communication and collaboration between 
groups, to make sure that specs work well together.

* W3C has gone out of its way to hire new people who are committed to 
openness in creating specs; among others in different groups, both Mike 
Smith and I strongly believe in working in the open, and promote that in 
the HTML, WebApps, and SVG WGs.

Not all of these changes are unequivocally positive.  For example, 
though the WHATWG were insistent that the HTML WG be open to anyone who 
wished to join, I think that the group would have been more effective 
and more civil if there were smaller core HTML Working Group amd a 
larger HTML Interest Group, with coherent policies for collaboration, 
decisions, and an appeal process from the beginning; with a more diverse 
community than the WHATWG, we need more structure in place.  As another 
example, the SVG WG conducts all its technical work in public, including 
posting its raw telcon and F2F minutes on the public list; but this 
sometimes costs us considerably more effort when people misinterpret 
those minutes, and we have to spend time clarifying misperceptions.

So, unfettered openness is not always beneficial, and it does not come 
without costs.  But the majority of these changes were worth it, and I 
think this shows a remarkable record of adapting to the ethos of the 
"Open Web".  In my opinion, the W3C still has clear improvements that 
could be made, but I do think we've made good progress, and and we are 
still changing.

And it's not just the HTML WG that is working closely with the WHATWG... 
the WebApps WG has a large degree of overlap with WHATWG, for example, 
and to a lesser extent, so have a few other groups.  (WHATWG is a bit 
hostile to SVG, but the SVG WG does try to be responsive to technical 
points raised there.)


> Currently the W3C and the WHATWG have very different document
> licensing policies.  The WHATWG document license allows anybody to
> create a derivative specification; the W3C's does not.  While the
> HTML working group requested a change to the license:
> http://lists.w3.org/Archives/Public/public-html/2009Feb/0388.html ,
> the W3C has not yet made such a change.  I think many members of the
> WHATWG community feel strongly about this issue; I certainly do.

To keep the record clear, the W3C did do the legwork around this issue, 
and presented our membership with various options; many on the Team feel 
the same way you do, though I personally see both sides of the issue. 
But as you know, our membership expressed concerns about fragmentation, 
though we did change to a more open license.

And even the WHATWG is divided on the issue, as you are aware:  of the 
four browser vendors in the WHATWG, in the W3C poll, two of them voted 
to adopt the license you prefer, one abstained, and one voted against 
the license.  So, this issue doesn't seem to be as much of a dividing 
wedge between W3C and WHATWG as you claim.

Just as with WHATWG, speaking of W3C as a monolithic entity is an 
oversimplification... W3C is not only the W3C staff, but first and 
foremost our members and our active participants.  W3C decisions are 
informed closely by what our members and community have expressed.  I'm 
not trying to shift responsibility away from the W3C Team, because 
ultimately we do try to guide the organization in the direction we feel 
is most productive, but these decisions are not made in a vacuum.


> The current licensing situation means that the only practical way
> the WHATWG and W3C can work together on the same specification is if
> all of the text originates on the WHATWG side.  That seems like an
> odd definition of collaboration, and I think it's closely tied to a
> number of the other issues causing conflict in this group.

The nature of the collaboration is that feedback comes from both the W3C 
and WHATWG lists.  I don't think anyone objects to the text of the spec 
originating in the WHATWG version, per se... that's not the issue.  The 
issue is that the WHATWG version is different than the W3C version in 
technical details, when it would be better for the larger community if 
they were identical.

By any reasonable measure, the W3C, and in particular the HTML WG, seems 
to be converging and collaborating, but the WHATWG doesn't seem to be 
doing the same.  This seems ironic to me, since the WHATWG is by 
definition meant to be a more flexible group than the W3C (which for 
better or worse adheres to the "rule of law"), but that doesn't seem to 
be the case.  In my opinion, the WHATWG would do better if it were to 
behave a bit more consistently, and if it played to its strengths in 
being more adaptable to the needs of the broader community.

So, the question remains... could and will the WHATWG help reduce the 
confusion in the community by maintaining a spec that is identical in 
its technical details to the W3C HTML5 spec?  Hixie suggested offlist 
that I email the WHATWG list, which I've done, but I'm not sure how that 
group manages inter-group conflicts and liaisons.

(Sorry for the slow reply, and any subsequent slowness... I'm on 
vacation, and only have sporadic email access.)

Regards-
-Doug Schepers
W3C Team Contact, SVG and WebApps WGs

Received on Friday, 25 June 2010 10:06:20 UTC