David H. Hansson (@DHH) sparked the “TDD being dead” latest twitter storm. DHH is no stranger to controversy, and he certainly picked the right title for it. Of course, the responses followed immediately. and most of them fell directly into his trap.
Most people were defending TDD. How? By saying that Ruby on Rails (originally authored by DHH) was built so badly that TDD was bound to fail on it. Others said DHH simply doesn’t understand TDD, and he’s a bad influence on software newbies. And some cited their excellent record in TDD and how it all made sense and created exquisite code. They were worried of the impact the article will have on people who don’t understand enough about the topic.
There was a very small minority saying DHH is right, and that TDD was a fad that didn’t live to its promise. They also mentioned failed experiences. Too few voices.
Since I like TDD and promote it wherever I go, you might think I’d join the pro-TDDists.
Not this time.
TDD doesn’t always work
Anyone who says differently is selling something. So, let’s just ignore the fact that many of the TDD supporters live off TDD in some way (coaches, trainers, yours truly as well). They have an interest in TDD promoted as good, (much like SAFe), and this kind of publicity is bad publicity. But let’s put that aside.
The TDD proponents actually believe that TDD is the best tool for the job. And it many times is.
But surely not all of the time, and for all people. There is no silver bullet, remember?
It may be that with enough experience and acquired skill, we can apply TDD to many problems, with ease and success. Not always.
When I started doing TDD, I dismissed it quickly as good in theory, but not likely to work in real life. It didn’t work for me then.
Many people in the beginning think the same. Some of them keep working in TDD and see the value, again in some cases. Some drop it off completely.
It’s ok to experience bad TDD.
DHH was right about one more thing, though
Did you notice how quickly (minutes!!) the discussion became about DHH the person, his understanding and skills? Rather than making the discussions about where TDD helps, and where it doesn’t.
We’re really good at religious wars in the software world, aren’t we? Isn’t it ironic that agile always happen in context, but if someone decides to against the stream, and he’s suddenly and obviously he’s wrong?
From contextual ideas to instant dogma in minutes!
As Uncle Bob Martin said in his response (in the more important 2nd half, of the article), TDD is a good tool, and many have found it worked for them. And as he said, if other tools provide the benefits of TDD, go ahead with them.
I second that. Let me quote one of my favorite sentences in software (somehow related to Bob):
“We are uncovering better ways of developing
software by doing it and helping others do it.”
We’re not done uncovering yet. TDD is one item we found on our journey, there are others.