Re: editors comments in the W3C HTML5 specification

On Apr 14, 2011, at 2:01 AM, Lachlan Hunt wrote:

> On 2011-04-13 10:57, Steve Faulkner wrote:
>> I don not believe the W3C HTML5 specification is a notebook for Ian to
>> scribble his personal opinions about changes he disagrees with.
>> 
>> The W3C HTML5 specification is being littered with comments by the
>> HTML5 editor,, these seem to appear whenever a working group decison
>> does not go his way,  for example:
>> 
>> [1]
>> <!-- The following paragraph is not included in the WHATWG copy
>> because it is wrong. For example, content models are not syntax. It's
>> also unnecessary. What kinds of things are conformance requirements is
>> explained in the previous section, which talks about RFC 2119. -->
> 
> That comment relates to this note:
> 
>  "the conformance requirements for documents include syntax (the
>   <table> element is conforming as a child of <body>, but not as a
>   child ot <title>), and semantics (the <table> elements denotes a
>   multi-dimensional data table, not a piece of furniture)."
> 
> The note itself doesn't make sense.  The fact that a table can be a child of body, but not title, relates to the content model, not the syntax.  I'd suggest that the note just be removed, but I'm not sure why it got added if the first place.

If we're going to be pedantic... when you think of HTML as a language, and consider only valid HTML to be correct sentences in that language, then content model is an aspect of syntax. It's not lexical syntax, but it does affect which character sequences match the formal grammar of "valid HTML5". So far as I know, valid HTML5 is turing-decideable, so it is at least a Type-0 grammar and may even be lower on the Chomsky hierarchy.

That being said, if there are any wording inaccuracies or even just infelicities in the cited note, bug reports are always welcome.

Regards,
Maciej

Received on Thursday, 14 April 2011 09:40:08 UTC