This is a topic that came up in a vigorous conversation recently. The example we used was whether or not tests should be written for accessors. I was on the yes side and will expose my arguments here.

My first argument was yes because it has happened time and again to me that the wrong member was returned by the accessor. The counter argument was that code inspection would often reveal those issues.

My second argument was that the tests are not written only to validate the code at the time it is written but also for the future when changes happen. The counter argument being that in the mean time it is that much more code that needs to be maintained.

My final argument was that in any case it was simpler and cheaper to set as a policy that everything should be tested and then agree on exceptions during code reviews compared to debate for every piece of code whether it should be tested or not. That being said the policy might also be ‘test everything except accessors’.

I failed however to make the argument that I find the most useful: if is not worth testing, why is it worth writing ?

Your thoughts, arguments and counter arguments, as always, are very much welcomed.

This entry was posted on Friday, August 27th, 2010 at 6:48 am and is filed under Driven. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.