A 10 mile run requires effort but that does not mean it should be avoided.

A 10 mile run requires effort but that does not mean it should be avoided.

Implementing and using Test Driven Development (TDD) and other related test-first techniques can be difficult. There are a significant number of criticisms of these methods.  These criticisms fall into two camps: effort related and not full testing.

Effort-related criticisms reflect that all test-first techniques require an investment in time and effort.  The effort to embrace TDD, ATDD or BDD begins when teams learn the required philosophies and techniques.  Learning anything new requires an investment.  The question to ask is whether the investment in learning these techniques will pay off in a reasonable period of time?  While I do not think I am the perfect yardstick (I have both a development and testing background), however I learned enough of the BDD technique to be dangerous in less than a day.  The criticism is true, however I believe the impact is overstated.

The second effort-related criticism is that early in a development project developers might be required to create stubs (components that emulate parts of a system that have not been developed) or testing harnesses (code that holds created components together before the whole system exists). However stubs and harnesses can generally be reused as the system is built and enhanced when testing.  I have found that creating and keeping a decent library of stubs and harnesses generates good discussion of interfaces and reduces the number of “I can’t test that until integration” excuses.  Again true, but overstated.

The third effort-related criticism is that in order to effectively do TDD you need to have automated testing (inferred in this criticism is the effort, time and cost for test automation).  I have seen both TDD and ATDD done without automation . . . my brother also had a tooth pulled without anesthetic, I recommend neither.  Test automation is important to making TDD efficient.  Writing test tools, learning the tools and writing test scripts does take effort.  Again the criticism is true, however test automation makes sense (and requires someone to learn it) even if you are not doing TDD.

The final effort-related criticism is that TDD, ATDD and BDD is hard to learn.  I respectfully disagree.  TDD, ATDD and BDD are different concepts than many team members have been exposed before in career. Just because they are different does not mean they are difficult to learn.  This is most likely a reflection of a fear of change. Change is hard, especially if a team is successful.  I would suggest implementing TDD once your team or organization has become comfortable with Agile and has begun to implement test automation which makes “learning” TDD easier.

A second class of criticism of TDD, ATDD or BDD is that these techniques are not full testing. Organizations that decide that test-first techniques should replace all types of testing are generally in for a rude awakening. Security testing is just one example of overall testing that will still be required. TDD, ATDD or BDD are development methods that can be used to replace some testing, but not all types of testing.

Also in the full testing category of criticisms, TDD does not help teams to learn good testing. I agree that just writing and executing tests doesn’t teach testing.  What does lead people to learn to test is the discussion of how to test and gaining experience. Agile teams are built on interaction and collaboration, which provides a platform for growth.  Neither test-first nor test-last (waiting until coding done) leads a team to good testing. Therefore this criticism is true but not constrained to use of TDD.

Embracing TDD, ATDD or BDD will require effort to learn and implement. If you can’t afford the time and effort, wait until you can.  Embracing any of the test-first techniques will require that teams change their behavior. If you can’t spend the effort on organizational change management, wait until you can. Test automation is important for efficient TDD.  If you can’t buy or write testing tools I would still experiment with test-first, but recognize that sooner or later you will need to bite the bullet. Finally, TDD is generally not sufficient to replace all testing.  The criticisms of are mostly true but they are not sufficient to overwhelm the benefits of using these techniques.

Advertisements