Re: Workshop and meeting requirements

On 2014-05-15 04:04, Charles McCathie Nevile wrote:
> TL;DR: skip this email. I'll summarise my thoughts on the thread 
> tomorrow with an updated outline for what I will propose to the AB.
>
> On Tue, 13 May 2014 20:16:43 +0200, Wayne Carr 
> <wayne.carr@linux.intel.com> wrote:
>> On 2014-05-13 09:34, David Singer wrote:
>>> There are a lot of inter-related concepts at play here, and some of 
>>> them are in conflict.
>>>
>>> Even if the process formally says that decisions made at face to 
>>> face meetings are not final, there is strong social pressure — and 
>>> indeed, for meetings to be effective, there needs to be strong 
>>> social pressure — to allow the meeting to set a direction and move 
>>> ahead.  “Look, maybe you have a point, but we discussed this for two 
>>> days in Uttar Pradesh, and we need very convincing arguments to 
>>> re-open the discussion.”
>
> Certainly. If it was discussed for two days by those able to attend, 
> or call into the meeting for two days in Uttar Pradesh, and when the 
> proposed resolutions are distributed to the mailing list for 
> confirmation, and clearly fail to show rough consensus, the problem 
> isn't the process, it is that the working group doesn't agree.
>
>> (while it isn't now) I'd like that to be the rule, that decisions at 
>> meetings have to be confirmed on the mail list (if only by object to 
>> decisions in the meeting minutes within 2 weeks or they stand). Given 
>> participation from all over the world, distributed meetings can be 
>> scheduled middle of the night always for some participants.
>
> Indeed. Or in the middle of a meeting at work that is very hard to 
> move, or...
>
>>> … a workshop … really ought to be possible to plan 8 weeks in advance
>
> I don't think this is a contested point.
>
>>> …prioritize attendance based on demonstrated interest and ability to 
>>> contribute (statements of interest, position papers, and so on).
>
> I don't see any argument that this should (continue to) be possible.
>
> I believe we need to clarify that this is not REQUIRED. It seems that 
> having it hidden among 4 other paragraphs in a chapter means people 
> don't look.
>
>> And a meeting that doesn't follow those rules could be called 
>> something other than a Workshop.
>
> I don't think that is a solution, it is just renaming a problem.
>
>> ... WGs can't have a WG meeting on short notice (unless there are no
>> objections),
>
> Which is important flexibility (leave aside that the 8 weeks is a 
> SHOULD for WGs anyway). Working Groups are meant to be working 
> together. One would expect them to be sufficiently functional that 
> where it is obviously important to get the group together on short 
> notice they can do so, and where participants have a real objection 
> this can be managed.
>
>> but they can have other meetings with the community or whomever which 
>> are not formal WG business meetings (e.g. cannot make WG decisions).
>
> This is currently completely undefined. The practice has been for 
> groups to assume they *cannot* do this, and therefore have "non-w3c" 
> events if they want something like a "workshop". In some cases it 
> should be easy to do this, in others not. I'll return to this question 
> below.
>
>>> But at the opposite extreme, sometimes a question comes up in a 
>>> working group, and a subset volunteer “look, let’s get together 
>>> around a whiteboard and in a place we can experiment, and we’ll 
>>> report back” that clearly doesn’t need nearly as much planning (and 
>>> it’s done all the time).
>
> There is an important key in this piece, which is that the subset 
> volunteers the information up front that they are doing this. At which 
> point the process of making it a WG meeting is usually trivial.
>
> There is nothing in the process that says how such notification should 
> take place - I think asking for a message to the mailing list defined 
> by the charter as the group's primary method for notification is not 
> onerous. In practice, it takes about 5 minutes to write such a 
> message, and is the sort of thing the chairs agree to do when they 
> take on the role.
>
> I think there is also a concrete difference between a specific-purpose 
> meeting (let's see if we can agree on good ways to make representative 
> decisions that will convince everyone else) and a more generic topic 
> like "let's figure out how everyone should do maths on the web".
>
>>>  Perhaps it’s worth stating as a principle that the strength of an 
>>> outcome is roughly proportional to the opportunity to be involved in 
>>> it; so a formal vote of all members of a WG is stronger than a 
>>> consensus at a face to face meeting, which is stronger than the 
>>> outcome of an informal get-together, and so on.
>
> Yes. I would have thought that was pretty obvious, at least to those 
> tasked with chairing a group - and therefore clarifying it for members 
> who don't understand that already. But perhaps it is worth stating 
> explicitly.
>
>> But informal get-togethers have no standing at all in the WG.
>
> If they really are informal get-togethers. It is trivial to hold a 
> meeting under W3C process - in essence, a WG says "we're having a 
> meeting" to somebody (their mailing list, or in a meeting, or to 
> someone walking down the street) and they have met the formal 
> requirements.

I don't think a Chair would get away with telling "someone walking down 
the street" would claiming that meets the requirements.  The process 
says "Meeting announcementsshouldbe sent to all appropriate group 
mailing lists, i.e., those most relevant to the anticipated meeting 
participants."  But, should means you need a valid reason and some 
thought about why you don't do it.  (otherwise there is no notification 
requirement at all)

But, getting back to your point.  Maybe there's some conflation of 
formal WG business meetings with other possible meetings of the WG. In 
the process, WGs can make decisions at WG meetings and there is the 
notice requirement.  That is a particular type of process defined 
meeting where formal business can be conducted.

I didn't think that precluded other gatherings of the WG, obeying all 
process requirements, where they cannot make decisions. e.g. they can go 
to dinner.  They can decide during a WG meeting at TPAC to offer a 
session to the larger community during the plenary day. They can decide 
last minute to attend some other WG meeting for an hour during TPAC.  Or 
they can decide to hold an informational gathering with the community 
outside W3C - like there's some sort of users group meeting in the city 
where the WG is meeting and the WG in their meeting decides to attend or 
send participants to that. I'd assume all of those gatherings of WG 
participants at meetings where they cannot conduct formal WG business 
would be fine and would not have any of the requirements of a formal WG 
meeting where decisions can be made.

The process needs to ensure that decisions are made fairly.  Any W3C 
related participation can't be in violation of anything in the process 
(or membership agreement or patent policy ...) regardless of whether 
that's in a particular type of meeting or not.

I'm not seeing why we need to define any exhaustive list of how WG 
participants can get together with one another or the community outside 
of their formal decision making WG meetings. The process document could 
have general rules ensuring transparency, participation, fairness etc 
that apply to any W3C organized gathering.  Does it need to anticipate 
every type of meeting W3C could come up with where no decisions are 
made?  (Workshops write reports that have an important justification in 
starting new work - so belong in the process document.  A workshop 
report because it follows those rules has more weight than a bunch of 
people getting together otherwise.)


>
> But there are benefits in following the process.
>
> The point about having meetings under the process is that there is a 
> legal difference between a standard developed in open collaboration, 
> and anti-competitive collusion. For companies in places where such law 
> is enforced, and for W3C which claims to operate under various 
> principles of fairness and transparency, this difference actually 
> matters.


I think this wandered a bit from the original point.  I don't think 
anyone was suggesting in this part of the thread that WGs should make 
decisions outside the process.

I was reacting to "a formal vote of all members of a WG is stronger than 
a consensus at a face to face meeting, which is stronger than the 
outcome of an informal get-together, and so on. "  Something like that 
would be worth capturing in a future version of the process (which seems 
a good idea, adding CfC, director appeal, etc).  But for that use 
"informal get-together" wouldn't be in the list of degree of 
establishing consensus at all.   It's not one of the ways of 
establishing consensus in the WG.  That's what I was trying to point out.

>
> There are also obvious benefits in the process being reasonable (which 
> IMHO it clearly is). And benefits in people knowing that - which IMHO 
> they clearly don't, by and large.
>
>> CfC doesn't seem to be in the process.
>
> There is a whole section on consensus. It boils down to something like 
> your suggestion below. But there isn't a defined rule for how they 
> will be done. Groups like the HTML Accessibility Task Force have 
> therefore explicitly agreed to and documented how they will implement 
> them [2]
>
>>  That may be worth adding with some description of how WG 
>> participants can request stronger forms of evaluating consensus.  
>> e.g. Director's appeal > mail list vote > mail list CfC > other means 
>> of Chair estimation of consensus (including decisions at WG meetings).
>>
> ...
>>> If it’s worth holding a meeting, it’s worth keeping and publishing a 
>>> record (I doubt anyone disagrees with this, though people tend to 
>>> get lax after a meeting has happened and the excitement has passed).
>
> I think that this is a point we should reinforce.
>
>> A WG can also decide to do something like hold an open discussion 
>> with interested members of the community somewhere they choose to get 
>> feedback on some question and to call it a W3C X WG community 
>> meeting, can't they?
>
> Perhaps. For most concrete examples I suspect the answer is yes.
>
> There is an unresolved question about what is in scope, because what 
> working groups do is determined by their charter - and this is a 
> formal agreement with the AC as part of the relationship under which 
> we make licensing commitments to W3C technology.
>
>>  (not a formal WG meeting, no WG decisions possible).  If the TAG had 
>> wanted to call the recent meeting a W3C Tag community forum, is there 
>> anything that would have prevented that?
>
> Probably not. About the only thing that would cause objection is if 
> the TAG agrees to take on work outside its scope (for most WGs that is 
> defined in their charter which can be revised. The TAG has a charter 
> but is also defined under the Process - and right now those two 
> documents seem to be in conflict. But that's not a relevant problem 
> for this set of issues).
>
>> If a WG wants to create a subgroup to investigate something to
>> generate information, not decisions, those people can meet without
>> any set notice, can't they?
>
> In principle, they *should* have 8 weeks notice, although they can 
> waive that with the agreement of the WG.
>
> So if the WG are happy with a subgroup that meets whenever they all 
> agree to do so, that's fine under the current process.
>
>> There's nothing blocking any of that, is there?  I don't see why we'd 
>> need to address any of that in the process.
>
> This depends what the subgroup does. If they are working on fixing the 
> markup errors in a few specs, I doubt anyone cares. But if they are 
> making wholesale technical decisions about the shape of e.g. Service 
> Workers or Web Components, and they expect their subgroup decisions to 
> be close to binding on the Working Group at large, they *should* have 
> a bad time if the Working Group at large disagrees.
>
> cheers
>
> Chaals
>