Re: [TDD] Unit testing
- Steve, sounds like a good call.
Shameless plug for others out there:
The DoubleDawgDare videos at
a refactoring screencast which has the big point: first strip the code down
to the best signal you can get, killing all noise, then work in readily
revertable small chunks.
Have fun! This is my favorite part. I wish strangers would just give me
money to sit at home and tackle their legacy.
On Thu, Jan 7, 2010 at 1:59 PM, Steve Howell <showell30@...> wrote:
> Good stuff. I think I am dealing with a slightly different kind of
> "interglobulationfulness" than your example, but it is making me think
> harder about how to restructure my code.
> I'm not really struggling with the typical problems of how to interact with
> a GUI, because my input mechanisms are fairly simple, and my output
> mechanisms are fairly easy to mock. It is more the relationships between the
> presenters themselves that is causing me consternation. So, in a way,
> creating more presenters (e.g. subpresenters) will only increase my problems
> in a way. As I look at my code, I am starting to think that I maybe I should
> first focus on easy problems, like coming up with better names for some
> variables, tightening up the view API slightly, and then come back to the
> harder problems.
> Steve Howell
[Non-text portions of this message have been removed]
> --- Steve Howell <showell30@...> wrote:--- Adam Sroka <adam.sroka@...> wrote:
>> You're still thinking too small! We need zettaTests, or
>> even...dare I say it...YOTTAtests!
> No. We need smaller. My yoctotests are a quintillionth theThat's it! If we want to standardize our test names, then I propose that we use a real standard:
> size of your microtests.
My "femto-tests" are smaller than yours: They tests individual quantum fluctuations within bits in the code under test (rather than lines, branches or conditions). My "mega-tests" check to see if the universe exists. With the Large Hadron Collider firing up, I figure we need to keep an eye on that one.