W3C home > Mailing lists > Public > www-qa-wg@w3.org > April 2003

leftovers on "normative" issue

From: Lofton Henderson <lofton@rockynet.com>
Date: Fri, 11 Apr 2003 09:56:32 -0600
Message-Id: <5.1.0.14.2.20030411092416.034ef080@terminal.rockynet.com>
To: www-qa-wg@w3.org

The minutes (excerpt below) say that we will start with the "what's 
normative" issue (LC-36, 65, 106, 108) Monday, specifically looking at 
response to 108.

So far, we have decided that the glossary is not normative (#36), and that 
we will work on section 3.1's specification of what's normative in 
SpecGL.  Specifically, improve the wording, and expand the bullet list as 
appropriate.  But NOT do section-by-section labelling in the body of SpecGL.

I think that we have generally agreed also that most (all) of section 1 is 
not normative, e.g., Jonathan Marsh asserted that the "priorities" 
definitions must certainly be normative (LC-106), but we believe that they 
are not.

IMO, the problem comes up because everyone has their own concept of 
"normative".  For example, here are SpecGL and UAAG definitions:

>SpecGL definitions [1]:
>-----
>
>informative text:  text in a specification whose purpose is informational 
>or assistive in the understanding or use of the specification, and which 
>contains no test assertions or conformance requirements.
>
>normative text: text in a specification which is prescriptive or contains 
>conformance requirements.
>
>UAAG glossary [2]:
>-----
>
>normative, informative:  What is identified as "normative" is required for 
>conformance (noting that one may conform in a variety of well-defined ways 
>to this document). What is identified as "informative" (sometimes, 
>"non-normative") is never required for conformance."

Our definition is more narrowly focused, and I think contains a notion of 
"directly connected" to conformance -- a test assertion or conformance 
requirement.   I think UAAG's is a bit looser, and admits to the 
interpretation that glossary definitions affect the meaning of conformance 
statements, and therefore are normative.  Similarly, for Marsh's assertion 
that our sec1.7 Priorities definitions *must* certainly be normative -- 
they are an aspect the mechanism by which we (SpecGL) measure conformance, 
but they do not directly relate to specific test assertions or conformance 
requirements of SpecGL.

This has some potential to become a "religious" argument, because of many 
different concepts of what is defined and within the scope of "normative."

Proposal.  I think we can clarify SpecGL and forestall future arguments, 
including commentors' rejection by of our rejection of their 
comments.  Here are a couple of ideas (in addition to what we have already 
decided):

1.) occurrences of "normative" and "informative" in sec3.1 should link to 
their definitions in Section 4 ("Definitions");

2.) the definitions in Section 4 should include a little discussion.  At 
the least, clarify and point out that our definitions mean "directly 
connected" by our definition; and, things like glossary definitions 
indirectly affect normative text, therefore are not considered normative 
under our definition; and, maybe point out that different WGs are using 
different definitions.

Regards,
-Lofton.


[1] http://www.w3.org/TR/2003/WD-qaframe-spec-20030210/#definitions
[2] http://www.w3.org/TR/UAAG10/glossary.html#def-normative

At 09:34 AM 4/11/03 -0400, Sandra Martinez wrote:

>QA Working Group Teleconference
>Thursday, 10-April-2003
>[...]
>LC-36, 65, 106, 108
>
>Comment about "Overall": Is the glossary normative?
>
>
>LR:     It is unclear which sections, paragraphs, and sentences are normative.
>
>LH:     What about Ian question. Is the glossary normative?
>
>LR:     Section 4 or the glossary?
>
>LH:     Both. See his e-mail he sent an example of normative in glossary.
>LR:     I consider the definitions in Section 4 normative. It Will be 
>helpful if we follow the Accessibility GLs and label all sections, as is 
>normative or not.
>
>LH:     What about 1.1 or priorities are they normative? Take the 
>difficult ones. I agree with you it is a good idea, but we will take a lot 
>of time agreeing, just take the example of priority. It is not trivial.
>
>LR:     Make sense. Can we just label some of them?
>
>PC:     Not a good idea. All comes down to the language.
>LR:     Going back to section 4. Is it normative? Maybe our definition of 
>normative is wrong. We need a new definition.
>
>PC:     I believe Section 4 is informative
>
>MS:     By definition glossary is informative.
>LR:     We just decided the glossary and the definition informative 
>sections are both informative.
>
>LH:     We should generate an action for someone to write Ian to ask for 
>comment on our resolution for this issue.
>
>LR:     We are comfortable with the definition section.
>
>LH:     I am not entirely convinced. We need to get buy-off.
>LR:     If we go to Ian as you suggest, we will be giving special 
>treatment, which implies going back to all the commentor.
>
>LH:     This one is more controversial.
>
>MS:     We should be consistent. I agree with Lynne, this could go on forever.
>
>KD:     As part of the process in resolving the issues, We need to go to 
>each individual and have a dialog with them regarding the resolution of 
>the comments and their agreement with the resolution.
>
>DH:     Agreed. Is better to have each individual agree.
>
>LH:     Do we have a response to 108?
>
>LR:     No response for 108 yet.
>
>LH:     We should start with 108 in our next meeting.
Received on Friday, 11 April 2003 11:54:53 GMT

This archive was generated by hypermail 2.2.0 + w3c-0.30 : Thursday, 9 June 2005 12:13:13 GMT