Re: HTML-AAM

Hi Rich - I put some thoughts on the ARIA list as Léonie had suggested:

https://lists.w3.org/Archives/Public/public-aria/2016Aug/0269.html

I meant to reply on this thread saying I had done so, but seem not to 
have, sorry.

Beyond what's in that message, some comments inline below:


On 30/08/2016 2:03 PM, Rich Schwerdtfeger wrote:
> Thanks Leonie,
>
> I can appreciate some of the overhead concerns, whereby the SVG A11y 
> task force needs to seek approval from within the task force, and then 
> by the SVG working group, and then by the ARIA working group. This can 
> take 2-3 weeks by itself.
Yes, the procedure to get TF approval followed by WG approval in 
sequence does slow things down. But it would be possible to do those 
approvals in parallel if the facilitators / editors were confident of 
their readiness.

In particular I want to note that the HTML AAM is not subject to this. 
We are not developing it in a joint task force, there aren't 
facilitators gumming up the works, regular teleconferences demanding 
peoples' time, etc. There are times when we think a joint TF is needed, 
but it's not a requirement and we decided it wasn't needed for HTML-AAM. 
The editors can (and I think do) seek publication approval from the 
sponsoring WGs in parallel, so there is no time cost to having to ask 
two WGs. Unless of course one of the WGs raises concerns, but in that 
case we're well advised to attend to them, and having the spec not joint 
wouldn't prevent that.
> Also, at the moment each of our groups works on a different process. 
> You work solely off of github whereas we have a combination of those 
> and will not switch entirely to github until ARIA 1.1 completes. This 
> change is to be more agile.
I think the ARIA WG is happy to adopt tools that lead to the most agile 
results. We use some tools that Web Platform doesn't (or in different 
relative weights), but for good reasons. AFAIK the HTML-AAM is being 
developed exclusively using tools preferred by the Web Platform WG, ARIA 
is not imposing other tools on them. We would continue to be open to 
other suggestions. We might not decide to have the ARIA WG use, for all 
its work, only the tools used by the Web Platform WG, but I think would 
be open to do so for the joint deliverable.

Another "tool" the ARIA WG uses is an editors' caucus, with a mailing 
list and twice-monthly meetings. Though the HTML-AAM editors do not 
participate in the meetings, I believe they follow the discussions, and 
certainly they benefit from the coordination efforts (for instance, the 
availability and maintenance of the mapping tables script, scripts that 
automate some common tasks, improved GitHub usage practices, etc.). I 
occasionally ask the editors if they want to continue to spend time with 
this forum and the answer has so far always been yes, so people are 
finding value. Nonetheless, we have not imposed specific expectations on 
degree of participation from the HTML-AAM editors, so I don't see this 
as burdensome, and net positive because of the gains that come out of it.
>
> I think it is extremely important, going forward, that platform owners 
> take ownership of accessibility infrastructure even if it means use of 
> specifications outside the working group (in this case ARIA). We also 
> need to be more agile and many of us appreciate the desire to not 
> duplicate process.
I agree to platform owners taking ownership, being more agile, and not 
duplicating process. I just think joint ownership doesn't compromise any 
of those things - it represents joint interest, shared process, and can 
be as agile as anything can be within W3C Process.
>
> If the Web Platforms working group would like to have sole ownership, 
> I would like to propose a one month review period for reviewing the 
> HTML-AAM CR spec. and quarterly updates of working drafts to ensure 
> that the ARIA Working Group, and others as well, be able to review the 
> specification.
I agree that this is important if Web Platforms were to be a sole 
developer of the spec, but think it would slow down the process further 
compared to the option of developing jointly, especially at the present 
state where the HTML-AAM 1.0 version is close to compete afaik.
>
> We should also work on the supposition of a good faith that both 
> working groups will do the right thing. I see no reason to think 
> otherwise.
I agree there is good faith from both WGs. But things change, and even 
if the WGs continue to operate in good faith, available resources and 
relative priorities can shift. Having joint ownership builds in more 
buffer to adapt to those changes over time.

Michael
>
> Michael, I would like to hear your thoughts?
>
> Rich
>
>
> Rich Schwerdtfeger
>
>
>
>
>> On Aug 30, 2016, at 12:12 PM, Léonie Watson <tink@tink.uk 
>> <mailto:tink@tink.uk>> wrote:
>>
>> On 30/08/2016 14:20, Rich Schwerdtfeger wrote:
>>> At the last meeting we agreed to look at taking this out for a CFC
>>> (HTML-AAM) ownership. A number of us, like myself, did not have an issue
>>> with this as we have always had a stellar working relationship with the
>>> HTML-AAM team in Web Platforms. Furthermore, I think there is value in
>>> host platforms taking greater responsibility for the accessibility of
>>> their platform.
>>
>> Thanks Rich.
>>
>>>
>>> That said, there were a some who had 2 questions which warrant a
>>> response that I cannot give them.
>>>
>>> 1. What is the tremendous overhead you are referring to? It takes about
>>> 2-3 weeks to get sign off on other specs. we have. … IOW it gets signed
>>> off by both working groups.
>>
>> Things are simpler when there is one parent WG. The work is done, 
>> issues are discussed, consensus is sought, and objections are handled 
>> in one place.
>>
>> Trying to do one thing in two places, especially where the two WGs 
>> have different work modes (as we do), often means that work is held 
>> up by one WG or another.
>>
>> IPR is less painful with one parent WG. It complicates things when an 
>> organisation has to consider IPR for a WG that has joint deliverables 
>> with another WG.
>>
>> For these and other reasons, the WebPlat chairs have a strong 
>> preference not to work on joint deliverables.
>>
>> Per the initial Github discussion [1], we are happy to become the 
>> parent WG and for ARIA WG to be named in our charter as an expected 
>> liaison. The proposed work mode would be to invite ARIA WG to provide 
>> feedback as the HTML AAM prepares to transition to CR (per the 
>> established wide review process), or more regularly if the ARIA WG 
>> would like.
>>
>> We would also be happy for the ARIA WG to become the parent WG for 
>> the HTML AAM, and for WebPlat to provide feedback as part of the wide 
>> review process.
>>
>> > 2. In the past we have seen things go out for review for a week 
>> which is
>> > not an adequate amount of time to review a spec. Given that there is a
>> > reliance on the ARIA spec. what is the review process you will put in
>> > place for external groups if we were to switch from joint 
>> publication to
>> > it solely being a deliverable of Web Platforms?
>>
>> We've already acknowledged that a week is insufficient time [2]. We 
>> asked in the initial discussion [3], and again on the ARIA call last 
>> week [4], how much time would work for you?
>>
>> Léonie.
>> [1] https://github.com/w3c/charter-html/issues/133
>> [2] https://github.com/w3c/charter-html/issues/133#issuecomment-242171664
>> [3] https://github.com/w3c/charter-html/issues/133#issuecomment-242171436
>> [4] https://www.w3.org/2016/08/25-aria-minutes.html
>>
>> -- 
>> @LeonieWatson tink.uk <http://tink.uk> Carpe diem
>>
>