FitNesse.
AcceptanceTests [add child]

Automated Acceptance Tests: Building the Right Code

FitNesse automated acceptance tests are power tools for fixing a broken requirements process. Skillfully applied, such tests make it possible to avoid the problems of ProjectDeathByRequirements. (Note: if you have not yet done so, you should probably first get a quick intro on FitNesse tests at the TwoMinuteExample.)

The Advantages of FitNesse Tests

FitNesse automated acceptance tests have several advantages over many kinds of traditional black-box, GUI-based functional tests:

Acceptance Tests vs. Unit Tests

If you are familiar with the emerging xUnit style of automated unit testing (e.g., using JUnit for Java, or NUnit for C#), you may well wonder why both kinds of testing are necessary, and how they work together. Let's discuss that.

xUnit: Building the Code Right

Automated unit tests (or programmer tests, as they are increasingly called), are white-box tests that describe and verify very low-level behavior. Especially when working test-first, programmers rely on such tests to make sure that the design is sound, and that there are no defects. Such tests are built by and largely for programmers, and can be independent of functionality.

If a team using JUnit or NUnit to do Test-Driven Development is still hampered by a broken requirements process, they can easily end up building robust, lean, well-factored code with no business value. (This is, in fact, happening with increasing frequency.)

FitNesse: Building the Right Code

FitNesse automated tests are about building the right code in the first place: the code with the most business value. While programmers may be involved in writing such tests, they are not just for programmers. FitNesse tests are especially for customers, business analysts, testing professionals, support staff, management, and all other stakeholders who need evidence that the project is producing solid return on investment.

"Running, Tested Features"

Ron Jeffries articulates the idea that the ultimate software project metric, the one closest to ROI, is Running, Tested Features. You can read about it here. How many RTF did we deliver last iteration, or last month, or last release? Only a tool like FitNesse can answer these questions.

Without a solid tool for automated acceptance tests, you cannot know how many RTF you produce per unit time, so perhaps you cannot easily determine what your project ROI is. Perhaps your team is headed for ProjectDeathByRequirements.

Consider Using Both Tools

If you use FitNesse without using JUnit or NUnit to test-drive and refactor your code, you may end up with tremendous business value, but rotten, buggy code. Even if the bugs don't kill you, this is short-sighted; it's a way of sacrificing release 3 to release 1. You have to keep the code lean and extensible to prevent code rot. Otherwise, your rotten codebase will need to be rewritten from scratch, long before it has exhausted its business value.

With both tools, you can balance production and production capacity. You can deliver the features that produce highest ROI, while producing robust, clean, extensible code. You might find that you get addicted to the early, frequent feedback that automated tests provide. Your whole team could get addicted to Building the Right Code, and Building the Code Right.

Summary

Learning More

If you have not yet done so, check out the TwoMinuteExample. For more information on how teams can work together with FitNesse tests to deliver better systems, see DeliveringTheRightSystem.

If you are ready to start creating FitNesse pages and test tables, check out EditingFitNessePages and CreatingTestTables.