...one major advantage of TDD is the ability to answer the question "how will I know I am done?". We can aswer that question because we have a test that will fail if we are not done and pass if we are.Of course this is exactly what I was on about when I wrote about why I like the Continuous Testing plug-in so much. (Thanks to Jason Yip for the pointer.)
Posted by Jeffrey Fredrick at October 19, 2005 08:58 PM
TrackBack URL for this entry:
I am curious... what part of Alan's claim is "bizare" as you say in your opening sentence?
Posted by: Antony Marcano on October 20, 2005 09:16 AM
By bizare I mean "bizare in the current climate", which is to say unfashionable. I'm really just echoing the point he makes in his entry which is that "It's very much in vogue at the moment.. (to say) of course, TDD is not about testing, it's about X".
Posted by: Jeffrey Fredrick on October 20, 2005 09:38 AM
If TDD is about testing, then it's about testing poorly.
I support TDD; I admire TDD; I *do* TDD when I develop test automation frameworks. But as a testing specialist, I worry about developers who are so crazy for TDD they lose all interest in the aspects of testing that are not captured in teeny-tiny unit tests written by people who have never studied and probably will never study the craft of software testing.
Posted by: james bach on October 20, 2005 01:20 PM
...Indeed, it is not in vogue to say so and I commend Alan for standing up and saying:
"TDD isn't *just* about Testing [...]"
It needed to be said! Let's hope that it is a view that spreads! :-)
Posted by: Antony Marcano on October 21, 2005 08:36 AM
James, two thoughts:
1. how can they lose interest they never had? my experience is that interest in the craft of software testing is rare even among testers let alone developers.
2. as a former tester myself I agree there is more to testing than unit tests, but just because a form of testing isn't complete isn't to say it is poor. that is TDD is about testing well... as far as it goes.
Posted by: Jeffrey Fredrick on October 21, 2005 01:29 PM
Let me add another note to what I take to be in "violent agreement" on this thread. If there was a way to measure testing goodness and design goodness in common units, then an average developer who practices TDD may produce 10 units of design goodness for every unit of testing goodness (I am just guessing the ratio here). This typically happens because it takes a much smaller number of tests to drive a testable design than to cover all of the combinations that need to be tested in a given piece of code. I actually don't care what TDD is about as long as I get to work with people who do it :)
Posted by: David Vydra on October 25, 2005 01:36 PM