Oh, sure -- but there are other tests for that. In particular, the BasicConversationTest -- a fairly involved scenario that I use as the primary smoketest -- tests comments going back and forth.

This one's specifically a regression test, so it's intentionally stripped-down. My philosophy of regression tests is that you should write a minimal test that clearly exemplifies the bug without additional complexities, then fix the bug, then re-run the test to prove that it fixed it as you expected...