Re: Plan/proposal for DoV group -- Monday

A couple of clarifications and proposal refinements are in-line...

At 10:08 AM 4/20/03 -0400, Lynne Rosenthal wrote:

>Comments in line
>
>
>>====================
>>THE PROPOSAL
>>====================
>>
>>Add a new Chapter 2, "Concepts", which might look like this:
>>
>>2. Concepts
>>2.1 Two kinds of guidelines
>>-----
>>[Restore lost bits from end of 1.7 in 20021108 version,
>>http://www.w3.org/TR/2002/WD-qaframe-spec-20021108/ ,
>>which said...]
>>
>>"The guidelines are of two general types:
>>
>>     * those that deal solely with the document features and conventions 
>> of the specification — GL1 and GL10 - GL14;
>>     * those that, in addition to documentation aspects, deal with how 
>> specifications should establish and define the conformance policy for 
>> the specification's technology, including ways in which the technology 
>> may be subdivided for conformance purposes — GL2 - GL9."
>
>This is in the current document, Section 1.5, Understanding and using this 
>document.

Ah, so it is.  But ... if we follow through with moving 1.8 into new 
section 2, maybe that also should be moved.  If you read all of 1.5, it 
stands out as a different in nature and purpose from the rest.



>>2.2 Dimensions of variability (DoV)
>>-----
>>[Move current 1.8 to here,
>>http://www.w3.org/TR/2003/WD-qaframe-spec-20030210/#variability]
>
>Yes.  The current text may benefit from some rewriting to improve its 
>clarity and understandability.
>
>
>>2.3 Specification category and class of product
>>-----
>>[To be discussed -- Per 20030418 telecon, it may be that we want to put 
>>the fully fleshed-out and improved discussion of SP and CoP in here, in 
>>their own subsections, with anchors on the two lists.  I think it is a 
>>good idea, as there were several comments about confusion about these 
>>concepts.]
>
>Yes.  The information is getting lost in G2.  This would also ensure a 
>listing in the ToC, which will help in finding this information.
>
>
>>2.4 Profiles, Modules, Levels
>>-----
>>[To be discussed (later, when we resume the prof/mod/lev issue?) -- It 
>>may be that we want to put a full discussion of profiles, modules, levels 
>>in here, with anchors on the two lists.  I think it is a good idea, as 
>>there were several comments about confusion about these concepts, "having 
>>trouble seeing a sharp distinction".  I don't think a clear picture of 
>>the concepts can necessarily be developed with very terse text, for 
>>reasons that I'll explain later.  We could take the time and space to do 
>>it here, perhaps recovering some discussion and verbiage that was in 
>>earlier drafts, but that has been dropped.]
>
>Yes. Having it all in one place may make the differences and similarities 
>clearer.  Also, as was suggested a diagram (if we can compose one) would 
>be helpful.
>
>
>>2.X (Sections for other DoV?  Or are they simple enough?)
>>-----
>>[ To be discussed --
>>
>>http://www.w3.org/QA/WG/lc-issues#x77 ,
>>http://www.w3.org/QA/WG/2003/03/DIWGcomments.html#SG-1
>>
>>seem to suggest that GL2, and 4/5/6 are the ones that need attention.  So 
>>maybe no additional 2.x sections are needed?]
>
>
>I'm not sure what you are thinking here.

In suggesting subsections for Ch.2, I responded only to commentor issues 
that highlighted a concept as confusing.  So I didn't propose a (ch.2) 
"Concepts" subsection on every DoV.  E.g., I didn't propose a subsection on 
"Discretion" or "Extensibility" (there were lots of comments on 
extensibility, but none based on "too confusing -- clarify"... in fact, one 
said "very well written").

So the question here is:

Alt.1:  do we stick with the topics and subsections that I enumerated 
already?
Alt.2: or, do we be comprehensive and have a subsection on every DoV?
Alt.3:  or, the enumerated subsections plus some (but not all) others?

>I think this may be covered by the above proposed sections.

Sounds like a vote for Alt.1?



>>2.Y Addressing relationships among DoV used
>>-----
>>[There is confusion about "address the interrelationships...":
>>http://www.w3.org/QA/WG/lc-issues#x21
>>http://www.w3.org/QA/WG/lc-issues#x99
>>
>>Here would be an opportunity to discuss the concept in general, perhaps 
>>with a generalized or generic version of:
>>http://lists.w3.org/Archives/Public/www-qa-wg/2003Apr/0121.html .
>>This discussion could then be back-linked from each of the relevant CP, 
>>which would then only add its own specialization to the discussion.]
>
>Good idea to address this.  I know that I am also confused about the 
>"address the interrelationships..." and what is needed to satisfy these 
>checkpoints.
>
>
>>GL2 Identify what needs to conform and how.
>>-----
>>[To be discussed -- the idea to move the detailed verbiage into 
>>"Concepts" (see above), and have abbreviated discussion here, heavy on 
>>links and references.]
>>
>>GL4/5/6 Profiles/modules/levels
>>[To be discussed -- the idea to move the detailed verbiage sections into 
>>"Concepts" (see above), and have abbreviated discussion in the GLs' 
>>verbiage, heavy on links and references.]
>>
>>GL2-9
>>-----
>>[Add a caveat back in.  For each DoV, we used to have a statement like 
>>the last paragraph of the verbiage at GL3 profiles:
>>
>>http://www.w3.org/TR/2002/WD-qaframe-spec-20020826/#b2b3b3d239 ]
>
>If we agree to have the caveat,  put it in the Concepts section and not in 
>each GL.   My concern is that dividing the technology is a good thing and 
>the caveat (although it says 'Excessive variability fragments') may give 
>the wrong impression.  Do others read it, as it was intended?

I don't know.  The caveats were basically put in there to satisfy the Dan 
Connolly comment, (non-LC) issue #69, and a months-long discussion about 
whether we could quantify how many DoV is too many.  Hence the parts of the 
caveat about the risk (combinatorial badness) of combining multiple DoV.

I still favor having something in place in each DoV.  But I could support 
something like this.

1.) Flesh out the current discussion in 1.8 (soon to be in Ch.2) a little 
better, put an anchor on it, and maybe even put it in its own sub-subsection.

2.) In each DoV GL, have something like "Exercise caution.  <BLAH> is one 
of the Dimensions of Variability.  See @@discussion@@ of risks versus 
potential benefits of using one or multiple DoV in a specification."



>>CP2.4, 4.3, 5.2, 6.1, 7.3, 8.5, 9.7
>>-----
>>[The "DoV relationship" CPs.  Their discussion/rationale would back-link 
>>to 2.Y, and maybe add some specialized discussion.]
>>
>>====================
>>ADDRESSING THE ISSUES
>>====================
>>
>>21
>>---
>>http://www.w3.org/QA/WG/lc-issues#x21
>>
>>This is solved by 2.Y above (and CP2.4, 4.3, 5.2, 6.1, 7.3, 8.5, 9.7).
>>
>>66
>>---
>>http://www.w3.org/QA/WG/lc-issues#x66
>>http://lists.w3.org/Archives/Public/www-qa-wg/2003Apr/0118.html
>>
>>This is not solved by the PROPOSAL (yet), but the above email proposal 
>>has gotten no negative comments and apparently stands.  To be discussed 
>>-- its explanation could be put somewhere in new Ch.2 (e.g., in 
>>prof/mod/lev, if such a subsection is approved).
>
>Perhaps the explanation could be put in ExTech as an example of what we 
>don't mean.

No strong feeling.  But a brief *positive* statement about the scope of 
DoV  -- as opposed to or in addition to the "example of what we don't mean" 
-- could be derived from this discussion, and would help to pre-empt the 
question in the future.

Your comment reminds me of ... the "ExTech List".  Is there one?



>>75.3
>>---
>>http://www.w3.org/QA/WG/lc-issues#x75
>>
>>The PROPOSAL does not address consolidating the DoV guidelines, 
>>GL2-9.  IMO, the only real possibility here, within reach for SpecGL 1.0, 
>>is the prof/mod/lev topic:  can those concepts be consolidated? or if 
>>concepts are kept separate, at least can the 7 checkpoints (4+2+1) at 
>>least be put under a single GL?
>>
>>Resolving that latter question must wait for the prof/mod/lev discussion 
>>(4/25?).  But on Monday we could endorse the idea (or NOT!) that that's 
>>the only potential opportunity for consolidation.

I meant 4/28 (we take it up on next Monday's telecon, after another week of 
completion of AIs and more email discussion.)


>>75.5
>>---
>>http://www.w3.org/QA/WG/lc-issues#x75
>>
>>This would be resolved by the GL2-9 ("Add a caveat back in...") 
>>comment.  Note the use of "a" -- it doesn't have to be exactly the old 
>>one, if someone has better words.
>>
>>84
>>---
>>http://www.w3.org/QA/WG/lc-issues#x84
>>
>>Solved by the PROPOSAL, I think.  The DoV are given prominence in Chapter 
>>2, and would show up in the TOC.  Ch.2 would link to the individual DoV, 
>>and vice-versa.  The DoV GLs and the Document GLs are not segregated into 
>>separate chapters (which would be the ultimate restructuring).  But IMO 
>>that is not needed.
>>
>>90
>>---
>>http://www.w3.org/QA/WG/lc-issues#x90
>>
>>Solved by the PROPOSAL.  The DoV are given prominence in Chapter 2, and 
>>would show up in the TOC.  Ch.2 would link to the individual DoV, and 
>>vice-versa.
>>
>>95
>>---
>>http://www.w3.org/QA/WG/lc-issues#x95
>>
>>NOT SOLVED YET.  If we agree that Conformance Policy is a DoV, then we 
>>need some verbiage.  I guess in GL3 would work, or another subsection in 
>>Ch.2.  To be discussed, 3 questions -- is it a DoV?  verbiage to answer 
>>originators question? if so, where to put it?
>
>I think the comment as merit.  Why is the Conformance Policy a DoV?  It is 
>a consequence of the other DoVs -- "It is important to convey an 
>understanding of what is meant by conformance and how it applies to each 
>class of product as well as each dimension of variability (e.g., modules) 
>if applicable."
>Hmm.  This should really say, "each other dimension of variability, since 
>CoP is a DoV.     Also, taking this literally, then the Conformance Policy 
>needs to be applied to itself, since the Conformance Policy it a DoV
>
>
>
>>77.SG-1
>>---
>>http://www.w3.org/QA/WG/lc-issues#x77
>>http://www.w3.org/QA/WG/2003/03/DIWGcomments.html#SG-1
>>
>>Resolved by 2.3, 2.4, and 2.Y.
>
>
>

Received on Monday, 21 April 2003 10:22:28 UTC