Re: HTML WG process

This starts almost the same as your other email for which I've not  
finished replying. I take it that this one supercedes? I'll act that  
way until informed otherwise.

On 22 Jun 2009, at 16:08, Larry Masinter wrote:

> I'm not interested in pursuing a legalistic
> review of the past and whether or not W3C
> HTML WG has or hasn't been in "violation"
> of the W3C process, because I don't think
> it would be helpful.

Then I would suggest that you don't make such claims. When I see such  
claims, qua member of the W3C community, I trying to investigate them  
(time permitting). Process violations can trigger large disruptions  
and can hurt mutual trust and smooth working whether or not they are  
"dealt with" formally. The W3C tends to promote its process as part  
of the benefit of working with the W3C, so I think that we should, as  
a community, take some care in using the charge of process violation  
(esp. when we don't intend anything to happen as a result) and try to  
substantiate such claims.

I don't think your claims stand up and you've just said you are  
unwilling to substantiate them. This leaves me a bit at a loss.

I accept without question that you do not like how the HTML WG is  
running.

> I didn't use the word "violation" in a formal
> sense;

It's an odd choice of word, then.

> the process document consists of guidelines,
> not laws.

I don't see how that's relevant as to whether you can violate it.  
Pretty clearly you can and certain consequences follow.

> I worked on the process document
> in the W3C Advisory Board,

I hope reasonable determinations of violation can be performed  
without having had to have been on the AB that drafted the document.  
That would make the process document of extremely limited utility.

> so I think I
> can speak to "intent".

Is intent relevant? Or strongly relevant?

I guess it is if you would say, now, "We meant that this situation  
would not be process legal but left a loophole in the language. Thus,  
I'm going to propose closing that loophole." But that's very much  
unlike your statement. In this scenario, the HTML WG is in compliance.

If you had said, "I think the HTML WG is only nominally in compliance  
with the charter while violating the spirit of it utterly," while I  
still think that's disputable, it's certainly less urgent a matter.

> For example, I'm pretty clear that the "SHOULD"
> indicates a requirement. When a SHOULD requirement
> is not followed, the reasons for not following it
> should be documented, clear, transparent and agreed.

Do you agree that violating a SHOULD requirement is less egregious  
than violating a MUST requirement?

The size of the working group, and the way decisions are made, are  
all documented, reasonably clear, transparent, and agreed (via the  
charter).

If your concern is about "timely responses" then I think the WG is  
obviously not in violation of this SHOULD requirement. It is pretty  
standard in the groups I've participated in to give a formal response  
to non-LC comments *when* the WG has resolved the issue. If the issue  
is contentious, that can take a while, but surely "timeliness" isn't  
a matter of a fixed time period (e.g., "Speedy trials" can take  
months or years to set up).

Perhaps you want that every comment gets an issue which can be  
tracked? And that's a "formal response"? Well, ok, but that level of  
detail is hardly mandated by the process.

> Otherwise, you risk situations where actors
> judge they are above judgment, assert their own
> criteria for neutrality, and corrupt the process.

Of course. On the other hand, I think neutrality and proportionality  
are important as well. I'm hard pressed to find textual and factual  
support that the "W3C HTML Working Group violates the W3C process in  
serious and significant ways". If the HTML WG *does* violate the  
process on this front, it's pretty clearly in at most minor ways,  
afaict. Furthermore, you aren't a neutral party in this having  
already expressed personal issues and, arguably, having institutional  
issues with the group.

This is why I asked for clarification from you. I don't think the  
HTML WG should be self judging (but I don't think that it is), but I  
also think that accusations of process violation have to be vetted  
carefully, esp. when it come from a source (yourself) who has great  
stature in the community (and can claim to have special expertise in  
W3C process).

> Personally the actual facts of the current
> HTML WG astounded me, were not clear before
> I joined the working group, only became clear
> after months in the WG, are not apparent from
> the W3C web pages describing the working group,
> and are inconsistent, in my opinion, with the
> intent and expectations of the process document.

As I said, I've observed the WG from the start. I've read every  
email. It all seems well within bounds with both the process document  
as written and with the practice of many other groups. It's better in  
some respects than some, worse in others, and unusual in the way that  
many groups are unusual.

I'll note that you seem to be conflating "intents and expectations of  
the process document" with *your* intents and expectations. This is  
not tenable, I hope you agree! Your bare claims can't be the basis of  
procedural assessment (other AB members from the time might have  
different views).

As with specs, we have to rely on the text for violation claims. For  
*sense* and *wisdom* claims, obviously, things are different.

To put it another way, I don't see how to use this claim when  
assessing future WG's process goodness without asking you. Indeed, I  
still don't know what your specific complaint is! It can't be  
"keeping an issue list of a certain form", as that's far too narrow.  
I'm really at a loss.

> Whether the difficulty is "serious and significant"
> is a judgment, but I think that should be determined
> by the effects will have on the goal of the process,

I think it's dangerous to conflate procedural correctness with  
outcome correctness.  They just aren't the same thing. It's  
important, dialectically, to be very clear about the distinction  
because it affects how people regard each other. Violating process is  
a kind of cheating, yes? Having bad outcomes isn't. People react  
differently to the different accusations.

> as I wrote a related post
> http://www.w3.org/QA/2009/05/_watching_the_google_io.html#c182444
>
> Whether the exceptions to the process discussed
> (and several others)

What are they?

> are or are not "violations"
> doesn't matter nearly as much as whether the
> abnormalities will cause continued and increasing
> fragmentation of the web.

I agree with this. Then why do you keep raising things as process  
violations? I would suggest you don't, so we won't be sidetracked.

After all, flip it around: Would you be any happier with the WG if it  
had the same outcome (fragmentation of the web) but was procedurally  
impeccable by your lights?

I monitor process claims because it matters to my organization that  
we understand the W3C process and can work effectively with it.  
(Well, also, I do care about procedural fairness, as well as outcome  
fairness.)

> Let's focus on the goal, and ensure that the process
> going forward is actually leading to convergence
> and interoperability. Increased transparency is
> only the first, but necessary, step.

I'm very much for transparency and openness. I don't think the HTML  
WG has been untransparant. On the contrary. While there is an  
"information overload" problem, that seems inevitable. I do see the  
WG trying to make that ever more manageable. It certainly is doing  
fairly well in comparison to many WGs existing and historical.  
(Compare any group that works on a closed mailing list. Compare with  
the complaints about getting minutes from the AB.)

So perhaps it all boils down to a rhetorical request: Please don't  
call something a *serious* process violation unless 1) you will  
substantiate it with the process document and 2) you really want a  
procedural resolution. It doesn't seem necessary to make your point  
and it risks engaging people (like me) on the *wrong* point (by your  
own lights).

Thanks for the reply.

Cheers,
Bijan.

Received on Monday, 22 June 2009 17:51:19 UTC