do-over on test approval

Today I re-opened our decision to approve

as it was evidently not clear to many WG members that this
involved approving the syntax
  SELECT ?name, ?mbox

Kendall apologized for being "asleep at the wheel" but actually,
he went on record as abstaining at the time

If more people had abstained, it would perhaps have been more
clear to the chair that the proposal should not be decided, but
rather, the discussion should continue.

Now it's not essential that everybody reads every character of
every test case, but if you're not going to read every proposed
test, you should think about who you're going to trust to do it for
you, ala "... hmm... Fred 2nded that proposal, and whenever
I play with his code I like what I see, so that's good enough
for me."

In the 30 Nov case, the proposal to approve the tests was
cited from the agenda, which was only about 1 hour short
of being available 24 hours in advance. In that case,
the "I haven't read them" excuse is pretty thin.
In the 7 Dec case, the agenda wasn't explicit about any
particular proposal to approve tests, so an "I haven't
read them" response is much more reasonable.

All this i-dotting and t-crossing arount test cases,
formal approval, not opening questions without new
information and so on may seem tedious, but I really
do believe it's a valuable service we provide to the
community in order to achieve a well-specified, well-tested
design, with a clearly-fair process behind it.

Dan Connolly, W3C
D3C2 887B 0F92 6005 C541  0875 0F91 96DE 6E52 C29E

Received on Tuesday, 14 December 2004 16:13:29 UTC