Comments on: Defining acceptance test criteria

By: andy

andy — Wed, 23 Jul 2008 15:20:20 +0000

@Vinny

The test is testing the balance and not the available funds.

By: Vinny

Vinny — Wed, 23 Jul 2008 09:27:18 +0000

I think Joes budgeting needs some work. If he has Â£1 and an overdraft of Â£100, his available funds are Â£101. If he withdraws Â£50, the balance of his available funds is Â£51.

By: andy

andy — Sun, 21 Jan 2007 12:45:08 +0000

I guess I did not explain that these are converted into executable regression tests.

By: Rob

Rob — Sun, 21 Jan 2007 02:07:37 +0000

I would caution that these be presented with a good dose of regression tests and standard test cases, lest users and managers become convinced that acceptance tests like these are a substitute for rigorous system testing.

By: andy

andy — Fri, 08 Dec 2006 21:02:55 +0000

I agree with you.

I also use lots of “Show and Tells” and encourage people to play with what we have built as much as possible.

It’s just that I like to have the conversation “how do I know this story has been completed to your satisfaction” so I have a sporting chance of catching the essence of what they care about. This is one technique that helps me do it.

By: Jason Yip

Jason Yip — Fri, 08 Dec 2006 20:31:37 +0000

I like this rigour and yet I am also wary that this approach can lead to people emphasising specification over empiricism. Is a system behaving correctly because all the acceptance criteria pass? Well, not if the acceptance criteria is wrong. How do we know the acceptance criteria was wrong? Well when we demonstrated it, we received negative feedback. When we tested it with users, they were not able to accomplish their goals.