Re: On what it means for a spec to be "normative" (re HTML5 & normative language spec)

On Sun, Jun 12, 2011 at 12:19 PM, Henry S. Thompson <ht@inf.ed.ac.uk> wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Jonathan Rees writes: ... I think this is
>> where your advice to write (non-normative) applicability statements
>> comes from.
>
> I fundamentally disagree.  It is a bad idea to think we are or should
> be in the business of trying to write tie-breaking rules

Sorry, "rules" is not at all what I meant - a rule would be normative,
and I excluded that. I certainly have no objection to what you say.

My "applicability statement" fantasy (not proposal) was of a statement
of intent regarding how to approach potential overlap or conflict - as
a guide for the perplexed, not containing any normative elements
("should" or "rule"). Maybe such a document would say exactly what you
have just said. Maybe concurrent publication by a single WG, or
normative reference of one document by another, implicitly constitutes
such a statement. Other cases are covered by introductory material in
actual specs, and so on.

I'm just riffing from first principles; I didn't mean to suggest any
particular approach to the situations at hand. I am struggling to
internalize and integrate a number of interesting ideas I've heard
recently and untangle some abstract concerns around agreement and
responsibility. I guess I SHOULD take the research off this list until
it's better baked.

Jonathan

Received on Sunday, 12 June 2011 14:52:36 UTC