Change § Testing and Debugging to use "fact" vs "assertion" terminology. about quantum-docs-pr HOT 4 CLOSED

@cgranade I agree, a few lines of clarification are certainly in order (it's getting late here, so I'll be taking care of it tomorrow). Changing the AssertPositive function name is obviously also necessary, but the question is to what? Perhaps TestPositive or FailIfNotPositive?

The clarifying discussion would then also serve as a good place to mention the actual Fact function in comparison to this to-be-renamed one---namely that this one actually halts the computation, whereas (it appears) Fact merely prints a message.

from quantum-docs-pr.

@gillenhaalb: Thanks for taking a look! With respect to Fact, it does actually halt if a particular fact is false (that is, Fact(true, "") is a tautology while Fact(false, "") is a contradiction). From that standpoint, AssertPositive could be renamed to PositivityFact.

The way I think of it is that a fact is either a tautology or a contradiction as a function only of its value: true is always true, irrespective of what the state of any control environment or qubit register is, while an assertion could be true or false depending on state external to the values passed.

The two can even be used well together, by representing global state with values that can then be used with facts. For example:

operation AssertQubitsAreAvailable() : Unit {
     Fact(GetQubitsAvailableToUse() > 0, "No qubits were actually available");
}

That couldn't be a function, since GetQubitsAvailableToUse depends on global state of the program and its execution environment (namely, how many qubits are available to use), but it can use that global state to find a value that can be used to formulate a fact.

from quantum-docs-pr.

@cgranade that makes a lot of sense, thanks! I think it may be worth adding that info to the API for Fact then, because it's hardly clear that it halts the computation (a feature which seems important enough to garner a few sentences of discussion on the Testing and Debugging page). It's only implicit in the description of the input string as "Failure message string to be printed ...".

I'll do an additional PR to the libraries to add the sentence in the function description.

from quantum-docs-pr.

Fixed by #695

from quantum-docs-pr.