Re: Conformance of SVG content in HTML

On Sun, 4 Oct 2009, Cameron McCormack wrote:
> Cameron McCormack:
> > > There doesn’t seem to be anything in HTML5 that defines the 
> > > conformance of an HTML document in terms of any foreign content it 
> > > contains. …
> 
> Ian Hickson:
> > Doesn't SVG already require that any SVG element be used in a manner 
> > that matches the SVG subtree concept?
> 
> It has a definition for what a conforming SVG DOM subtree is.  I still 
> think it would be more logical for the outer spec (HTML, in this case) 
> to specifically invoke the right conformance definition in SVG.  But 
> after discussing it in the SVG WG meeting last week we feel that 
> validator implementors are likely to do the right thing anyway (i.e., 
> check SVG subtrees for conformance) if they understand HTML5 and SVG, so 
> I’m not going to press the point.

I think considering HTML the "outer spec" here misses the point. It's just 
as possible for HTML to be inside SVG as for SVG to be inside HTML. Just 
as SVG doesn't have a clause for HTML conformance of HTML fragments inside 
<foreignObject> or MathML conformance of MathML fragments inside 
<foreignObject>, it would seem odd for HTML to have clauses for particular 
vocabularies inside HTML.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'