While I tend to agree that automated testing is in fact "checking," I think there is an element of getting hung up on words here. Glenford Myers' original book and the ISTQB certifcation people (which we may nor may not value) describe software testing as an action whose purpose is to find software defects.
If you broaden testing to more a scientific definition: executing a hypothesis regarding the expected behavior of the software, then regression testing is the act of verifying whether older hypotheses are still valid. It's only regression if we assume that all prior hypotheses are true, but regression can also tell you that former expectations are no longer valid because of recent changes.
While I tend to agree that automated testing is in fact "checking," I think there is an element of getting hung up on words here. Glenford Myers' original book and the ISTQB certifcation people (which we may nor may not value) describe software testing as an action whose purpose is to find software defects.
If you broaden testing to more a scientific definition: executing a hypothesis regarding the expected behavior of the software, then regression testing is the act of verifying whether older hypotheses are still valid. It's only regression if we assume that all prior hypotheses are true, but regression can also tell you that former expectations are no longer valid because of recent changes.
I agree, but I see this seldom executed like that in practice...
While a broader definition can be used to incorporate all of this "testing" I think that the distinction Boch is making is an important one.