Test reviews

[Robert Orosz <roboro@auto-trol.com>]

All,

In yesterday's working session we decided that each new test in the test
suite should be reviewed by someone other than the test author. Lofton
expressed a preference that the review is assigned to an implementer that is
planning to start implementing (soon) the functionality being tested by a
given test. We also decided that each reviewer should not have to review
more than three tests. This leads to a problem. There are 20 news tests in
need of review, and we only have four implementers. That means that only 12
tests will be reviewed by implementers leaving 8 tests still in need of
review.

One possibility would be to have each implementer review five tests instead
of three. How you would feel about that? The reviews would still be in an
area that you are actively working on to implement.

Stuart expressed an interest in doing reviews, but as he pointed out he
would be a bit in the dark since he wouldn't have access to an
implementation that supported the functionality being tested. I would be
willing to pitch in also with the same caveat that Stuart expressed, and I
believe Franck also expressed an interest in doing one or two reviews.
Perhaps the three of us could be secondary reviewers.

What are everyone thoughts on this?

Rob

<<application/ms-tnef>>