W3C home > Mailing lists > Public > www-qa@w3.org > February 2002

Re: Exit Criteria (CR/PR) Interoperability report

From: Al Gilman <asgilman@iamdigex.net>
Date: Mon, 25 Feb 2002 11:15:02 -0500
Message-Id: <200202251615.LAA203250@smtp1.mail.iamworld.net>
To: Karl Dubost <karl@w3.org>, "Ian B. Jacobs" <ij@w3.org>
Cc: www-qa@w3.org
Sorry to be such a broken record.  I just want to explain the rationale a little more.

It is all well and good that we have _defined_ the process in a pure logical form.

What we have to deal with is the fact that the readers of this document almost all have background in programming, and in programming, "exit criterion" from a loop is a very strong cultural artifact or shared concept.  This concept will grab anything vaguely close to it unless the circumstances require more precision, and in those cases we bear the burden of explaining the difference, motivating the use of some other less familiar concept.

Karl is not the last person who will stumble on this point.

Programmers will instinctively look at CR as an _activity_ which is carried out under a "DO .... UNTIL" control construct.  And the expression in the "until" clause is known as the exit critieria for this continuing activity.

Since this interpretation fits, to first order -- until one gets out the surgically sharp pencil, it is necesary to handle this distinction in a reader-friendly way, and not simply use our own internal definitive code language.

[Citations concerning "how to read W3C-speak" supressed out of laziness.]

Al

A quality glossary [process first, product following] does a compare and contrast with terms that are close in sense and in wider circulation than the term introduced.

At 10:01 AM 2002-02-25 , Karl Dubost wrote:
>
>
>KD: A word to add to the W3C glossary: Entrance Criteria and in the 
>definition, we may explain that we should not use exit criteria.
>

Make the binding more friendly.  Say "when people refer to exit criteria for Candidate Recommendation phase, they are generally referring to the criteria required for entry into Proposed Recommendation Status."

The point is that "CR-exit" should be treated as a colloquialism that is current in discussions, which we wish to capture and relate to the PR Entrance Criteria notion in the Process Document which should be used instead for precise work.  In the Glossary alone may not be sufficient.  This linkage may need to be expressed in the principal appearance of the notion in the body of the document as well as in the glossary.

[...]
>
>KD: To add to issues list for the QA WG. Remove CR-exit vocabulary.
>

See suggestion above for how to handle this.

Al

>
>>>Maybe we need a new checkpoint. Because it's one of the formal 
>>>thing written in the Process document but not yet clearly 
>>>explained. :)
>>
>>
>>IJ: I think it's explained very clearly. Read it without thinking 
>>about CR exit and I trust you will find that the process holds 
>>together. I don't think the description is currently broken, I think 
>>that people are used to talking about "exit" criteria even though 
>>the Process Document does not.
>
>
>KD: Ok. The process is clear on the notion of Entrance. There's still 
>a need for the checkpoint to explain how to do it. A checkpoint + 
>Techniques.
>
>Thanks Ian for your comments and to have pointed out the abuse on language.
>
>
>
>
>-- 
>Karl Dubost / W3C - Conformance Manager
>           http://www.w3.org/QA/
>
>      --- Be Strict To Be Cool! ---
> 
Received on Monday, 25 February 2002 11:15:15 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Sunday, 6 December 2009 12:13:58 GMT