On 09/04/14 08:26, John Mark Vandenberg wrote:
I agree with everything Risker said. I go further and
suggest the team
involved stops defending their goals and implementation. The former are not
the issue, and the latter was indefensible. I havent looked at how much
testing was done, or if there was some staging of the rollout, but it is
clear that it wasnt careful enough.
While I agree in general, I'm not even sure about that, that the goals
weren't the issue - do we even know what the goals /were/, or the issues
that they were trying to address? I've asked, and seen others ask,
several times in multiple venues, and I've yet to see a real answer. The
most anyone has been able to produce (usually Steven, I think, so I'll
thank him for at least trying) have been things like that it looks
better or it's more consistent, but that would be like a security person
only saying their change makes things more secure, and explaining
'because security' in the commit message. Sure, it might be true, but
it's completely useless if you're trying to work with it, or test it
exhaustively. We need to know what the issue really was, too, or we
can't possibly have any real discussion or collaboration. With security,
it's communicated - we just saw that with announcements of the
heartbleed exploit. Why isn't it with design?
So, yeah, the goals are probably noble enough, but we don't even know
that for sure. There's a gulf here, and it's just getting worse.