[wbs] response to 'Call for Review: Web Platform Working Group Charter'

The following answers have been successfully submitted to 'Call for Review:
Web Platform Working Group Charter' (Advisory Committee) for Vivliostyle
Inc. by Florian Rivoal.


The reviewer's organization does not support this Charter for the reasons
cited in comments but is not raising a Formal Objection.

Additional comments about the proposal:
   I have two main concerns about this charter. I have considered raising
them as formal objections, but decided against because these are long
standing issues that I do not expect to be resolved overnight, and I do not
wish to obstruct important work. Nonetheless, I think it is important for
the health of this working group, and of the W3C, that they be addressed.


== 1. Relationship with the WhatWG ==

This is an issue I raised against the previous charter, and I consider it
unresolved. https://github.com/w3c/charter-html/issues/112

A few of the documents maintained by the WG, most notably HTML, are also
maintained by another body, namely the WhatWG. The charter does call for
"mak[ing] an effort to avoid differences" between the two documents, but
while this is laudable, it is insufficient. The easiest way to avoid
differences is to not maintain the document in two places. Given that we
are not doing that, we must have reasons for maintaining these documents.
Any difference we introduce between our version and the whatwg version
needs to be viewed in light of these reasons.

Unfortunately, these reasons have not been written down, and we lack a
clear consensus. I believe this fundamentally undermines the mandate of the
working group, its chairs and its editors, as they have nothing clear to
back up their actions.

This is critical, because what the goals are and how we prioritize them may
completely change how we should work on these documents.

For example, if our highest priority goal is to make sure that the spec as
it is being implemented, is covered by the patent policy, then any
deviation from the whatwg, which is the one that implementors look at, is a
bug. In that context, converting to bikeshed, was a mistake, and doing
anything but treating the whatwg document as the editors draft is going to
be a source of difficulties. This does not mean that we cannot participate
in the discussion and influence the direction, but if pantent protection is
the primary goal, we need to optimize for the spec staying in sync.

On the other hand, if we believe that W3C style consensus-based decision
making, drawing from a broad and diverse membership, is a superior way to
make decisions and that making sure it applies to these specifications is
our primary goal, then we only need to consider the whatwg as one source of
input, and conversion to bikeshed or modularization are good things, given
that they make participation easier.

I do not expect this to be an easy question to answer, nor do I expect
everyone to agree about which goals are relevant or what their relative
priorities should be. But I fear that unless we tackle this, a growing
amount of people will disengage from the W3C side of the effort out,
leading to the worse possible outcome: a set of specifications that are not
followed by implementers (and therefore irrelevant from a patent point of
view), not the result of a broad-base consensus, yet having still borne all
the cost of developing them, including the ongoing political cost of our
not-so-healthy relationship with the whatwg.


== 2. Mandatory incubation ==

In the middle of a controversial debate about whether mandatory incubation
was a good idea, last year the WebPlatform WG was the first WG to include
it in its charter, making it an experiment worth following both for
supporters and detractors of the idea.

The debate is not closed, and people still fiercely disagree. Before making
the experiment permanent, and potentially using it as a model for other
groups to follow, I believe we need to take a step back and evaluate the
results first.

The AC was presented with a status update from the WICG, but it did not
really answer the questions. That the WICG is able to function is not
evidence that it is preferable for established WG to push their early /
experimental work to it.

I would like to see statistics about the number of specifications started
in the WICG. How many of them died? How many of them lived on? How many of
those that lived on have graduated to a WG versus how many are still in the
CG? Of those that are still in the CG, how many are there because they are
still exploratory versus how many have turned into living standards? Of
those that have graduated to a WG, what was the typical level of stability?
Of those that have died, what were the main causes for giving up? How good
was the progress along TR for the specs that did make it to a WG? Are all
the statistics roughly the same depending on who originated the idea, or
are there significant differences when the originator is a browser vendor
vs a non-browser but experienced spec writer vs a new-comer? How do these
statistics compare to when early/exploratory work was still being allowed
in the WG?

A well constructed study along these lines could go a long way to put to
rest the concerns that have been raised against mandatory incubation, or on
the contrary it may show that they were justified. Or maybe statistics is
the wrong way to go about it, but nonetheless, we need to draw lessons for
the life-sized experiment we've been running, and do so in a way that seems
fair to proponents and opponents to the idea.

I believe that renewing the charter with the mandatory incubation clause
as-is, without doing this introspection exercise is a mistake. It leaves
all the tensions and disagreements unresolved, which adds friction and
occasional conflicts to the daily work of the consortium and its members.


The reviewer's organization intends to participate in these groups:
   - Web Platform Working Group

The reviewer's organization:
   - intends to review drafts as they are published and send comments.

Answers to this questionnaire can be set and changed at
https://www.w3.org/2002/09/wbs/33280/webplatformwg-2016/ until 2016-09-30.

 Regards,

 The Automatic WBS Mailer

Received on Wednesday, 28 September 2016 05:51:09 UTC