Taking a Pragmatic View of Isolated Tests

Indeed! I mean the second one.

This is a case where a bedrock assumption has become unstated: I just don't write tests that depend on the sequence of execution any more. :)

I've added a footnote to try to clarify my meaning.