Re: TAG minutes 10 Sep for review (f2f planning, websockets URI scheme, HTML, ...)

> NM: the bit about "or other applicable specifications" is  
> potentially really cool... though it's not terribly clear. ... does  
> it mean "specs from the HTML WG"? or "from W3C"? or "from
> WHATWG?" or "Joe in his basement"? ... some responses in blogs  
> suggest the most liberal interpretation
>
> danc: a request for clarification of "applicable" would be  
> strengthened by inclusion of an example, such as "is spec X  
> applicable?"

"Other applicable specifications" means specifications recognized as  
applicable by someone applying the HTML5 spec to their activities.

This is already how things work with other specs:

For example, XForms 1.0 is defined by the W3C and it *could* be  
applied to extend XHTML 1.0. Yet, vendors shipping XHTML 1.0  
implementations in UAs don't support XForms (at least not by default)  
and even the W3C Validator doesn't bother to support XHTML 1.0 +  
XForms 1.0 validation. Thus, for a substantial part of the Web  
community, XForms 1.0 isn't "applicable" as far as XHTML 1.0 goes.

This shows that being published by the W3C doesn't automatically make  
a spec "applicable".

On the other hand, XHTML 1.1 is a closed set of elements and  
attributes. In particular, it doesn't contain RDFa attributes.  
However, to the community that recognizes RDFa in XHTML as legitimate,  
RDFa in XHTML is "applicable" to XHTML 1.1 and "XHTML 1.1 + RDFa 1.0"  
is ok for this community even though the RDFa attributes invalid in  
XHTML 1.1 alone.

This shows that what HTML5 makes explicit is already an implicit  
extension point in other specs.

As for "Joe in his basement", Mark Pilgrim specified feed  
autodiscovery for HTML and XHTML on his blog in a blog post. Multiple  
UA vendors implemented his spec and countless authors made content  
using this language feature. Feed autodiscovery is, therefore,  
"applicable" to the processing of HTML 4 and XHTML 1.0 as far as a  
sizable part of the Web community goes.

If you compare feed autodiscovery and XForms, it turns out that things  
specced by "Joe in his basement" can be more "applicable" than  things  
specified by the W3C. Therefore, applicability doesn't depend on who  
asserts applicability but on whether the community accepts the  
assertion.

For the purposes of Validator.nu, I've used a rough guideline of  
considering a spec applicable if two browser engines out of the set of  
Gecko, WebKit, Presto and Trident make a non-trivial effort to support  
a spec. All four made an effort to support ARIA before ARIA was  
officially integrated in HTML5, so Validator.nu had an HTML5+ARIA  
validation target with ARIA as an "other applicable specification"  
before ARIA was officially part of HTML5. Gecko, WebKit and Presto  
support SVG 1.1 in application/xhtml+xml, so Validator.nu supports SVG  
1.1 as an "other applicable specification" in XHTML5. However, for the  
time being, it doesn't support SVG 1.1 in text/html. Also, it doesn't  
support SVG 1.2 Tiny, because of the four engines, only Presto  
supports SVG 1.2 Tiny. MathML 2.0 is supported as an "other applicable  
specification" in XHTML5, because Gecko and Presto support MathML  
(well, the presentational part of MathML but I was too lazy to  
specifically axe the semantic part of the off-the-shelf schema as  
unapplicable).

-- 
Henri Sivonen
hsivonen@iki.fi
http://hsivonen.iki.fi/

Received on Thursday, 17 September 2009 07:54:54 UTC