Review of QA test guidelines


After our failure to submit a review of the QA specification guidelines,
I have sent a personal review of the QA test guidelines that are not yet in 
last call.

It is, if anything, more opinionated, and less helpful than my earlier 
opinionated and unhelpful review of the specification guidelines, and I am 
not seeking WG endorsement.

Document reviewed:


However, one thing I would point out is that in their framework:
- we would have test assertions scattered throughout S&AS. In particular 
every necessary and every sufficient condition would become an RFC 2119 MUST
- we would have metadata with each such assertion
- we would have at least one test for each assertion
- we would relate each test to every tested assertion

If anyone feels motivated I would welcome contributions of such tests or 
such metadata.

We certainly could do with more tests. I believe focussing on the issues 
for which we do not have tests is likely to be more cost effective.


Received on Tuesday, 1 July 2003 12:35:11 UTC