Re: [Issue-101] Replacement for the other of the diagrams in the Process document

On Thu, 08 Oct 2015 20:50:21 +0200, Stephen Zilles <szilles@adobe.com>
wrote:

> From: Jeff Jaffe [mailto:jeff@w3.org]

> On 10/8/2015 1:55 PM, Wayne Carr wrote:
>

>> On 2015-10-06 21:18, Jeff Jaffe wrote:

>>> On 10/6/2015 11:50 PM, Stephen Zilles wrote:

>>>> See below

>>>> From: Jeff Jaffe [mailto:jeff@w3.org]

>>>> Steve,
>>>>
>>>> In CR, if there are substantial changes we stay in CR; but if there  
>>>> are very substantial changes >>>>we go back to WD.
>>>>
>>>> Have we defined the difference between substantial changes and very  
>>>> substantial changes?
>>>>
>>>> [SZ] No, IMO it is up to the WG to decide whether the changes would  
>>>> take a document out of CR.

This is not correct. Revising a CR requires approval from the director, so  
attempting to change it in a way that did not satisfy the director - for  
example because there is no evidence that the changes proposed have  
received sufficient review - is what would effectively require a new WD.  
(The director may also explicitly require the group to go back to WD, but  
that seems a rarer case).

>>>> The most obvious reason would be that there need to be major
>>>> implementation changes and, therefore, the document is not really
>>>> ready for implementation (the anachronistic definition of CR)
>>>> anymore. If the changes involve implementation tweaks and the
>>>> participants agree that they should be made, then those are
>>>> substantive, but not “very substantive”.

The question is not whether the participants agree that the changes should  
be made, it is whether they can convince the director that the changes  
have been appropriately reviewed.

In terms of how this would work, a transition approval might suggest  
changes and say "the editors suggested this would fix a problem", the  
'director' reviewing the request might say "ORLY? And did they ask anyone  
else what that means", the staff contact would say "well, here is a thread  
where the WG discuss the proposal and agree it is great, here are 8  
outside implementors who independently noted the same problem and  
suggested a fix along the lines being proposed", and the Director would  
say "Hmm, makes sense…".

Alternatively the staff contact might be "we asked a few people and the  
internationalisation people said it made it impossible to use anything  
except ASCII to write people's addresses, and the accessibility people  
said it would no longer be possible to use the spec for someone without a  
mouse, and the privacy people said it exposes people's intimate life to an  
unconscionable degree, but we figured we'd like to fix the text. In which  
case presumably the director would refuse permission to publish the  
revised CR.

In terms of how we express that in a diagram, I understand the difficulty  
Steve faced in finding the right term, because it doesn't leap to my mind  
either. The rough scenario is "change that needs further review", but that  
scrimps on details and bulks up the image already...

>>>> If you think putting in some text like that would be useful, it could  
>>>> be proposed

>>> I wasn't making a proposal. I was just trying to make sense of the  
>>> terms in the proposed diagram. Your explanation doesn't make it clear  
>>> enough (at least for me).
["substantive change" has implications for revising a Rec in particular]
> The diagram differentiates between substantive change and very  
> substantive change.  I don't see any such definition in the process  
> document.
>
> Perhaps you are arguing that substantive change is "3 Corrections" and  
> very substantive is "4 New features". If that is the intent, I would  
> prefer to use that language in the diagram.

That would work for me as an approximation, although it isn't perfect  
either. But then, this is an illustrative diagram. The people who rely on  
it *without* reading the linked text (and in an update I'll fix the links  
so the labels really link to the right bit in the document - along with  
some necessary improvements to make the thing sufficiently accessible) are  
always going to have a bit of a bad time. They are also probably always  
going to exist.

> [SZ] Two points:
>
> 1.    The goal, posited by David Singer, was to have the labels on the  
> arrow indicate the trigger condition for the transition. For the  
> transition from PR back to CR, the trigger is “substantive
> change” I do not believe that we should change that label.

Agreed.

> 2.    For the transition from CR back to WD, I tried to create a label  
> which would reflect a trigger because the Process Document only says  
> that “Return to Working Draft” is a possible next
> step for a CR, but does not give any “trigger” for that step to be  
> taken. Given that observation, perhaps the label should be ,  
> “Substantive Change and Working Group Decision”, implying that it is up  
> to the WG to decide whether the document should go back to WD.

Well, the WG can decide that, but the trigger is "don't get permission to  
revise the CR in place, and don't want to go forward without the  
showstopping revision(s)".

> I do not see a requirement to go back to WD if new features are added.

There isn't one. Instead, a new exclusion opportunity arises. Assuming all  
the other requirements (dependencies met, enough review, etc) are  
satisfied.

> There is such a requirement for Edited Recommendations, but not for CR.  
> Perhaps there should be such a requirement.

I don't think there should be such a requirement. The point of it for  
Edited Rec is that if we're going to start bringing in obvious new IPR,  
it's worth the process of spinning up a formal set of patent commitments  
to it backed by the membership at large, rather than letting a rump group  
of 4 heroic editors just blithely assume the world is benign...

cheers

-- 
Charles McCathie Nevile - web standards - CTO Office, Yandex
    chaals@yandex-team.ru - - - Find more at http://yandex.com

Received on Thursday, 8 October 2015 23:10:57 UTC