W3C home > Mailing lists > Public > public-w3process@w3.org > October 2014

Re: 2014 Process: WD -> CR difficulties

From: Nigel Megitt <nigel.megitt@bbc.co.uk>
Date: Fri, 3 Oct 2014 16:04:47 +0000
To: "Michael Champion (MS OPEN TECH)" <Michael.Champion@microsoft.com>, "Wayne Carr" <wayne.carr@linux.intel.com>, Arthur Barstow <art.barstow@gmail.com>, public-w3process <public-w3process@w3.org>
Message-ID: <D0548532.1382C%nigel.megitt@bbc.co.uk>

Factoring into procedural vs declarative aspects makes a lot of sense: you
can't completely avoid having any process though, otherwise nothing will
actually happen. Similarly, you get the benefit of the declarative
approach if doing so avoids a lot of 'code' repetition. In a single
process with only a few steps it probably doesn't make much difference.
Either way the machine has to turn the 'code' into real world actions.

Anyhow, if the process were to disappear then outside observers would say
W3C is 'just a bunch of people publishing documents they think are pretty
neat for themselves' rather than 'a bunch of people going through a
rigorous process to ensure that the documents they publish are pretty
neat, for a lots of other people too'. The process adds a huge amount of

The point here is that the process needs to be clear both internally and
externally. If you refactor so there are lots of complicated lookups 'when
state X has been reached do action 1; state X is true when conditions a, b
and c; condition a is true when condition d, e, f' etc. that can reduce
clarity for everyone, even if from a documentation perspective it's more

Which is a long winded way of saying the process document has to get the
balance right. Profound? Er, no.


On 03/10/2014 16:32, "Michael Champion (MS OPEN TECH)"
<Michael.Champion@microsoft.com> wrote:

>> hat's not a process then, it's a set of what some call "stage gates"
>> with verification steps. The document should be renamed
>That's an interesting point, although I'm not sure there would be a net
>decrease in confusion if we renamed the "process" document.
>The "aha" moment for me in the long AB discussions around making  WGs
>more agile was when someone (Ian Jacobs maybe) made the analogy with the
>declarative vs procedural/imperative meta-discussion in computer science.
>For example 

>declarative-and-imperative-programming . Rather than trying to devise an
>algorithm that simultaneously a) ensures that relevant stakeholders have
>had a chance to review and comment on a spec; b) makes the W3C Process
>less opaque to non-specialists; c) allows WGs to be more agile.... it
>made sense to specify what a good result looks like and let WG figure out
>how to parallelize the steps and optimize the checks in a way that works
>for their community.
> Perhaps it is a misnomer to call a declarative definition of what a good
>result looks like a "process", I believe it really is the result we are
>interested in here, not certification that all the hoops were jumped
>through. The 2014 Process document doesn't go as far as it could in the
>"declarative" direction, but FWIW I would prefer to move it more in that
>direction than to add back "imperative" steps, whatever we end up calling
>the document itself.
>From: Nigel Megitt <nigel.megitt@bbc.co.uk>
>Sent: Friday, October 3, 2014 1:43 AM
>To: Michael Champion (MS OPEN TECH); Wayne Carr; Arthur Barstow;
>Subject: Re: 2014 Process: WD -> CR difficulties
>On 02/10/2014 17:51, "Michael Champion (MS OPEN TECH)"
><Michael.Champion@microsoft.com> wrote:
>>We can make the list public, but we can't make the necessary people
>>subscribe to the list.
>>The AB and/or the Process CG discussed this in some depth while
>>deliberating the 2014 process.  My recollection of the consensus was that
>>-- consistent with the spirit of the new process philosophy -- the
>>Process Document doesn't describe a machine that you crank to put out
>>Recommendations, it describes *what*criteria a spec must pass to become a
>>Recommendation.  *How* that happens can be optimized by specific WGs and
>>Chairs to work in the actual environment they live in.
>That's not a process then, it's a set of what some call "stage gates" with
>verification steps. The document should be renamed to avoid creating
>misleading expectations and make obvious the change in philosophy.
>Google "define: process"
>        process
>        noun
>        1. a series of actions or steps taken in order to achieve a
>>-----Original Message-----
>>From: Wayne Carr [mailto:wayne.carr@linux.intel.com]
>>Sent: Thursday, October 2, 2014 9:43 AM
>>To: Arthur Barstow; public-w3process
>>Cc: Nigel Megitt
>>Subject: Re: 2014 Process: WD -> CR difficulties
>>On 2014-10-02 04:30, Arthur Barstow wrote:
>>> On 10/1/14 3:21 PM, Wayne Carr wrote:
>>>> There could be a Call for Review public mail list.
>>> Agree [and it might even be useful if the `right` people subscribe
>>> In case you did not know, the [chairs] list is already used to: 1)
>>> make FPWD transition requests [rarely do these fail]; 2) announce LC
>>> publications + explicit  RfC from specific group(s); 3) make CR
>>> transition requests. It would be helpful (vis-à-vis toward getting
>>> early and wide review) if all three of these (plus ProcDoc-2014 now
>>> effectively mandates a "RfC for pre-CRs") were announced on a Public
>>> list.
>>> Unfortunately, the chairs list is Member-confidential and I suspect
>>> subscriber membership is controlled by consortium staff (i.e. I don't
>>> think it is an  auto-subscribe-able list by Joe Public). I would be
>>> delighted if everything on that list was automagically forwarded to a
>>> Public list. However, I suspect typical Public vs. Member
>>> confidentiality stop energy would prevent that :-(.
>>Can W3C staff just make this list?
>>public-wg-call-for-review@w3.org or public-wg-rfc@w3c.org
>>- posts should only be from WG Chairs and W3C Staff
>>- notices:
>>    + Staff sends announcements at: transition requests, "last call"
>>publications - FPWD and the CRs that have substantive changes or Last
>>Call (under the old process), notice that work is underway on a charter
>>    + WG Chairs send RfC on anything the WG would like feedback on or
>>that they would like to tell the public.  e.g. want review on a
>>particular section, notice that a section is considered stable
>>This could just be done and WGs use it as they see fit and the process
>>could later mandate it (if that was wanted).  In the meantime, it would
>>be something WGs and W3C staff could use as a way of asking for reviews
>>or making general announcements on spec development.
>>> (WRT `the tools will save us`, if WG charter deliverables included
>>> some type of "interestedGroup" property, then it seems like at least
>>> some notifications could be automated.)
>>> -AB
>>> [chairs] https://lists.w3.org/Archives/Member/chairs/


Received on Friday, 3 October 2014 16:05:24 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:35:12 UTC