[whatwg] [html5] tags, elements and generated DOM

Fri Feb 24 15:06:54 PST 2006

On Thu, 7 Apr 2005, Henri Sivonen wrote:
> 
> If you mean: Is it worth going down the road of requiring getElementById 
> to support arbitrary characters in ids?
> 
> Perhaps not. Depends on what is implemented.

I expect the spec to (when I get to gEBI) say that all characters must be 
supported. It's easier to implement, for one.

> If you mean: Is it worth going down the road of requiring getElementById 
> to consider idness based on factors other that the ID type in a DTD?
> 
> Definitely yes. Browsers don't process DTDs and the W3C already allows 
> the DOM impl to use info other than the DTD to figure out which 
> attributes count as ids.

Yeah, that's already water long under the bridge.

On Thu, 7 Apr 2005, Henri Sivonen wrote:
> 
> I don't think SGML validation is part of What WG conformance 
> requirements. I thought Hixie has specifically said he doesn't bother 
> with DTDs.

Indeed. WHATWG will happily host schemas and DTDs if people write them 
(see, e.g., syntax.whatwg.org; fantasai maintains that). However, none of 
them will be given any sort of preferential status. They're just 
implementations.

On Thu, 7 Apr 2005, Henri Sivonen wrote:
> 
> I am very hostile towards the idea of requiring UAs to implement any XML 
> parsing features that are in the realm of the XML 1.0 spec but that the 
> XML 1.0 spec does not require. This means processing the DTD beyond 
> checking the internal subset for well-formedness.
> 
> I would rather suggest that What WG specs explicitly discourage people 
> from using a doctype on the XHTML side and point out that authors should 
> not expect UAs to process the DTD.
> 
> Those who want to use entities for input, should parse and reserialize 
> as UTF-8 in their own lair and not expose their entity references (or 
> parochial legacy encodings) to the public network.

The spec has text to this effect in places now; let me know if you have 
more specific text you'd like to see. I don't want to be too strong, since 
if you're using XML, exactly how you do so is the problem of the XML spec, 
not the Web Apps / XHTML5 spec.

On Thu, 7 Apr 2005, Henri Sivonen wrote:
> 
> Ideally, UAs would know nothing of that particular doctype and would 
> trigger the standards mode because there is a doctype that is not on the 
> list of doctypes that triggers the quirks mode or the almost standards 
> mode.

Currently the spec says you trigger quirks mode if there isn't a <!DOCTYPE 
HTML>, and you feel like it (both conditions must be met). I may make this 
more specific in due course; do people think we should?

I'm also tempted to just codify many parts of quirks mode; do people want 
to do that instead?

> > Or at the very least use something that would not confuse people into 
> > thinking that it is an application of SGML or XML.
> 
> Do you want to replace "NONSGML" with "THIS-IS-NOT-SGML"?

Well, we're now down to just <!DOCTYPE HTML>, since that was the shortest 
I could make it while still triggering strict mode (well, <!DOCTYPEHTML> 
worked too, but that just looks silly).

Suggestions on shortening it are welcome.

I really wish we could ditch the DOCTYPE altogether. The number of pages 
that use quirks mode is just sad.

On Fri, 8 Apr 2005, Henri Sivonen wrote:
> On Apr 8, 2005, at 11:05, Jim Ley wrote:
> 
> > The proposed string that MUST appear as the first line of a WHAT-WG 
> > document is... please do not call it a doctype unless it is a doctype, 
> > see even people on the list are confused by using this!
> 
> Well, it could be defined as a tag soup construct called "doctype", 
> which is neither an SGML doctype nor an XML doctype. :-)

Indeed, it's a DOCTYPE token, and it generates (in the right place) a 
DOCTYPE DOM node. It's not an XML or SGML DOCTYPE per se...

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