AcceptanceTests

This is the new approved name for functional tests. It is better because it describes their purpose: to show that the application is acceptable to the customer. Would someone describe their experience with acceptance tests here because they have been very weak in the projects I have worked on ? ChrisCottee³host³³date³August 15, 2000³agent³Mozilla/4.0 (compatible; MSIE 5.0; Windows NT; QuickBooks? 6.0 UK; DigExt)³AcceptanceTests

We've gone really hard at acceptance tests for the past couple of years. I don't think your comment is unusual: in fact I think most of the XP community has little idea how to do this. One major reason has been the lack of a widely accepted framework in the style of xUnit for unit testing.

So here's some pointers:

Ward Cunningham's FiT framework is about a year old now and marks a welcome shift towards the community taking Acceptance testing more seriously. The framework is nice to use but is heavily numerical (seems mainly suited to testing applications whose output is single numbers, such as financial applications) and is something of an extension of unit testing approaches, in that it essentially verifies the domain classes rather than the application itself. WillemVanDenEnde also makes this point in his comments on FiT under AcceptanceTesting.

Then there's my own TextTest framework. Which is, naturally, God's gift to acceptance testing frameworks ;) We have, at least, built a very successful XP project on it with an astonishingly low bug count (see AcceptanceTestOnlyDevelopment), and it has been much easier to persuade other projects to use it within my own company than to get them to adopt XP full-on. It took the step out into the open source world in June and is now available from SourceForge, and has been successfully adopted on a few projects at Astra Zeneca now, where they, unlike us, combine it with a unit test suite.

-- GeoffBache