Re: Report on event listener investigation

On Aug 28, 2014, at 7:56 AM, Janina Sajka <janina@rednote.net> wrote:

> …let me clarify that my concern isn't so much protecting the
> reputation of any individual, but rather the PF as a whole, and
> particularly the process. While I would not want us to have to closely
> guard what we say, words matter; and what we say here is archived.

I am sorry, and apologize, if you or anyone else found what I wrote offensive. It was not intended to be so. I am also interested in protecting the reputation of the PF Working Group. I believe its work is valuable. Otherwise, I would not be a member.

> So, for the sake of the record it's important to put on the record that
> we're not seeking to get around anyone or any W3C group even as we try
> to define precisely what we're looking for and how we think it might be
> achieved. That's not a case of circumventing W3C process or groups,
> though it might legitimately postpone particular conversations.

As you said earlier, we are discussing options:

>>>>> I see Shane's email discussing various documented, legitimate
>>>>> approaches. Extension specs are one such approach agreed by consensus of
>>>>> the HTML-WG as part of Plan 2014. It's not circumvention to suggest PF
>>>>> might want to propose moving forward via an extension specification.

One of the options listed was to have PF or HTML-A11Y-TF develop an extension specification for a new property on the EventTarget interface, exposing the currently registered event listeners. My feeling is that this is not a good direction to follow: not because it's procedurally inappropriate (it is appropriate procedurally, we are clearly allowed to develop extension specifications), but for two other reasons:

* DOM specification development really needs the expertise of people who do 
  that; this is a specialized area with interesting dependencies and cross-
  links. There ARE areas where extension specifications make sense — where 
  what they specify is essentially independent, but this is not one of those 
  cases.

* Significant, and almost certainly predominant, use of a listeners extension
  to the EventTarget interface, if it is specified, would be for non-
  accessibility reasons and use cases. This means that an accessibility group 
  would either have to gather and understand and consider all these use-cases
  outside our area, or we would risk getting it wrong for some other 
  constituencies.

The other specialist considerations for this interface that lie outside the expertise of PF might include (and I say might, because I am not myself a DOM designer, and I just see the possibility of issues or questions in these areas):

  a. potential for abuse,
  b. security,
  c. user privacy,
  d. rendering engine performance, and
  e. interaction and overlap with other DOM interfaces.

I'm offering personal opinions on the options here; my feeling is that pursuing this option would not work well for the overall web and W3C community, for the reasons above. I hope this helps to move the discussion along.

Cheers,
James

Received on Thursday, 28 August 2014 20:10:19 UTC