Re: Request for clarification on HTML 5 publication status (ISSUE-19)

Julian Reschke wrote:
> Not really. What I was asking for was a *normative* spec that defines 
> the language, not the DOM, not error recovery, not client-side storage etc.

Any spec that defines the language should be specifying:

1)  How to parse it.
2)  How to parse strings that claim to be in the language but are not.

Of course an "acceptable" specification for #2 is "throw up a big error message 
and crash"... but is that really desirable?