Re: keep conformance objective (detailed review of section 1. Introduction)

Charles McCathieNevile wrote:

> If we want it to be a must, we can make it so. What Dan is asking is 
> that we agree to your statement and make it a principle.

I think perhaps James Graham's observation is the most apposite :

> James Graham wrote:

>> If we require a term for documents that pass the subset of conformance 
>> constraints that are machine checkable, may I suggest "valid". The draft 
>> defines a "HTML5 validator" as a mechanical conformance checker for 
>> HTML5 documents. The term "valid" is also in common usage to indicate 
>> that a document has passed automated conformance checking.

There are very good (pragmatic and theoretical) grounds
for differentiating between "valid" and "conforming", and
I would be very happy to tone down my own statement to
refer to "validity" rather than "conformance".  The set of
"valid" documents are those that pass formal validation;
the set of "conforming" documents are those that (a) are
valid, and (b) "conform" to other requirements that are
not amenable to machine verification (such as, for example,
the requirement that <blockquote> contain a quotation, and
be not simply a convenience for achieving visually indented
text).

Philip TAYLOR

Received on Tuesday, 25 September 2007 09:44:59 UTC