Should? Really?

[MockOfT]

IMO:

I mean it in the nicest way possible (although it is going to be hard).

1. For me to understand what the tests do, I look for "Is" or "Are" or "Must", "HasTo", "Does", "Calls" but not "Should"; not this soft and nice "Should"; it sounds like "if you have a chance to do it" or similar.

2. Having it as an extension and using it at the end of the statement is harder to read and understand the intension. One is Asserting that something works and therefore stating it in the first place.

[ipashchuk]

@MockOfT , raising this issue is basically saying that this library is flawed to its core... I and tens of thousands of developers who are using this library respectfully disagree.

1. In legalese, "should" or "shall" is a contractual obligation, not merely a suggestion. But, I understand your point and agree to some extent that "should" has much less force than "must" in our vernacular. There is an interesting discussion on this point here: https://softwareengineering.stackexchange.com/questions/45783/when-writing-specs-bdd-style-should-you-use-should-or-should-you-not.
2. I completely disagree. I have a result, I want to see if it's as expected. I would much rather assert:
   actual should be equal to expected
   than
   assert equality between actual and expected
   nUnit, for example, easily allows developers to transpose actual and expected parameters