[rspec-users] Spec'ing exhaustively?

David Chelimsky dchelimsky at gmail.com
Sun Aug 24 11:07:56 EDT 2008


On Sun, Aug 24, 2008 at 9:44 AM, Muhammad Ichsan <ichsan at gmail.com> wrote:
> Dear All,
>
> Before I know spec'ing, I've been practicing testing using JUnit. I
> remember that once Kent Beck said (if I'm not mistaken) that we should
> not test exhaustively. Instead, we should test until our doubt in the
> quality is gone. What about spec'ing? Is spec'ing aimed to create good
> documentations or just to remove the doubt? If it is to remove the
> doubt, then spec'ing will be pragmatic not exhaustively.

Spec'ing, or "coding by example" as we've been calling it lately, all
started with TDD. No matter what you call it, or how you approach the
process, the goal is the same: reliable, maintainable software.
Testing "until our doubt in the quality is gone" makes sense to me
regardless of whether you call the process TDD, BDD, testing,
spec'ing, coding by example, etc, etc, etc.

I include "maintainable" in my definition of quality. For me, if code
examples document the code and that documentation makes it easier to
maintain, than those code examples serve to increase confidence (or
remove doubt, if you prefer) in quality.

Make sense?


More information about the rspec-users mailing list