The stuff i use a lot gets refactored a lot until i am happy with the design.
Actually, Bret refactors to the point where I find his code a little difficult to parse in my head. :)
One thing at a time.
Amen. In other words, I agree with everything Chris said. The only problem I
have with Kevin's article is that it is written somewhat from the perspective of developer-testing. I would love to see a similar article directed more on testers. Perhaps Chris can write it.
That's why every time I plug this article I mention that it bears re-reading. Read it once and you'll get a some ideas to use. Use them. Then read it again, and you'll get more ideas. Repeat until you've finally figured out what he was actually saying. He hasn't written that much about it, but Jonathan Kohl is also really good at this kind of larger-scale test automation design. He's probably a better candidate than me to write the Definitive Article on the subject. One thing we pretty much all agree on is the agile formula: make it work; then make it good; then make it fast. Your first pass is not going to be pretty. That's OK. My first pass is procedural; Bret's first pass is crazy OOP or something; I don't know what Jonathan's first pass is like, but I gather it has a lot of code duplication. However you can get it to work. The point being that lots of testers are in your boat, and information about your situation is floating around unorganized at the moment. But it's coming, slowly. You're on the right track. Keep going, keep asking questions, and keep us posted.
_______________________________________________ Wtr-general mailing list [email protected] http://rubyforge.org/mailman/listinfo/wtr-general
