sábado, abril 01, 2006




> The question I have is, at some point in my process the method I am
> working on was "good enough" to add unit tests; I delayed adding tests
> until the code was "really good". Is this something you find yourself
> doing? How does it work out for you?

I often have the same feeling as you: after doing a bunch of unsafe
refactorings (refactorings unsupported by tests), I have the code
good enough that I feel that I can understand it and modify it. I'm
tempted not to write the tests.

Like many temptations, this is probably not a good idea. I think it
was Michael himself who said that if you don't put some tests in
today, the system will still be untested a year from now. Presumably
I was here to make a change: I owe it to myself, and to the
universe, to do enough testing to show that first, the change is
needed (red bar), then that I've done it correctly (green bar).

I urge myself to press on, and sometimes I have the foral mortitude
to actually do it. I hope you're stronger, and better, than I am.


aahahaahahaahahahahAHahahAHaHaHAHA

see that's the main difference between OOP and electronics/physics/math related stuff, OOP is human... it fights, it laughs, it is a constant battle between propose and functionality. There's no way you can say A+B=X in OOP, all you know is that A operation B =X, there are a gzilion ways to define this operation, you have to find the better one: safer, faster, simpler.



Sem comentários :