Re: TWO Change proposals for ISSUE-41 : Distributed Extensibility

"Robert Ennals" <> wrote:

> Referring to the original ISSUE-41 definition, I think what people
> want is the ability to define extensions similar to SVG, MathML, and
> FBML that can be used in an HTML document.

In order to evaluate your proposals, I think it would help if people who've previously wanted Distributed Extensibility to be added to HTML5 commented on whether your proposal X and Y constitute Distributed Extensibility in their opinion. (Even better if they could formulate criteria that could be applied to determine if a Change Proposal would introduce Distributed Extensibility into HTML5.)

Shelley doesn't agree with your proposal ( Tony prefers Proposal Y (

I'd be interesting in learning the opinions of the TAG members and Sam's opinion given their involvement is raising ISSUE-41 in the tracker.

About your concrete examples:

I believe adding SVG and MathML into text/html the way they have currently been added to HTML5 yields a superior authoring experience compared to using either Proposal X or Proposal Y to add SVG or MathML to text/html. I would, therefore, consider SVG and MathML to be anti-examples of why Distributed Extensibility in the form of either Proposal X or Proposal Y would be desirable.

As for FBML, I note that I've never seen a Facebook rep ask for Distributed Extensibility in these discussion. I realize that the existence of FBML could be taken as an unvoiced need, though. 

Proposal X isn't quite clear on how names with a colon in the markup would be exposed in the DOM. However, the first bullet point under Negative Effects suggests that a colon-separated prefix in the tag name would affect the namespaceURI property of the corresponding element node. If this is the case, implementing Proposal X in Gecko would break the actual FBML unless Facebook cooperated and changed their implementation. (Facebook's FBML library has distinct code paths for "IE", "Mozilla", "Opera", "WebKit" and "Other". If a browser is sniffed to be on the "Mozilla" code path, getElementsByTagNameNS(document.body.namespaceURI, "fb:foobar") is expected to match element nodes whose tags looked like <fb:fooBar>.)

> It is often useful for people to define extensions to HTML. These may
> be vendor-specific experiments for features that may eventually get
> folded into the main HTML spec.

How can the prefix ever be gotten rid of once deployed?

Henri Sivonen

Received on Thursday, 18 March 2010 14:28:18 UTC