W3C home > Mailing lists > Public > www-tag@w3.org > June 2011

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

From: Tim Berners-Lee <timbl@w3.org>
Date: Mon, 13 Jun 2011 16:07:01 +0100
Cc: Jonathan Rees <jar@creativecommons.org>, Larry Masinter <masinter@adobe.com>, "www-tag@w3.org" <www-tag@w3.org>
Message-Id: <D7632962-ED64-4A85-904C-787005BFC663@w3.org>
To: Karl Dubost <karld@opera.com>
Yes, in the use I am used to it sn't that a spec is normative, but that in a given
spec, there is text which is or is not normative and references which are and are
not normative.

To meet a specification, you typically have to obey the normative bits
while being informed by the non-normative bits, and you generally
have to also implement the obey the object of references.

Presumably to say that a spec itself is Normative could mean that
(a) it is formally written with normative bits so can be used as a normative spec by others
or (b) is in a given context, like a common mother spec, a direct or
indirect normative reference.
I don't think W3C has a concept of a Normative Spec in the 
absolute, unless I've missed it.   In the HTML5 case perhaps people would mean
a normative reference, direct or indirect, from the HTML spec.

Tim
Received on Monday, 13 June 2011 17:42:15 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:48:35 GMT