Forum Overview :: NetHack
 
Re: Caltrops, please share your experiences with BDD or ATDD by E. L. Koba 04/17/2013, 9:45pm PDT
I've been doing full-on BDD for the last 2+ years with some absolute BDD believers, and I've come to the conclusion that the theory is totally flawed, but the outcome is not bad.

In short (because I just got back from work/drinking):

Your product guys are never going to see any benefits to writing their tickets your new arbitrary format (I'm working with the best product guys I've ever worked with and they always roll their eyes whenever we start talking about cucumber)

The cuke's that non-devs write are just going to be totally re-written by devs to match their existing step definitions.

Nobody except the devs are ever going to look back at your existing cukes - so really what is the point of writing them.


Fullofkittens wrote:

I've kinda been tasked with figuring out how to bridge the gap between our nontechnical QA and our devs so that we can increase the scope of our test automation and release more often.

To that end I'm giving a talk tomorrow about BDD frameworks (Cucumber and SpecFlow) with the goal of trying to use the BDD frameworks as a point of collaboration where the nontechnical QA can at least participate and describe what they want automated and the devs can make it happen.

I would rather see QA learn how to automate shit on their own but they've been kind of resistant to the idea of... learning things.

Any experiences?


Writing good cukes and steps is probably harder than writing the code. You end up going down one of two paths, both of which kind of suck (but could be improved with better tooling):

1) All of your cucumber steps are general and now you've created essentially your own new DSL (which everyone needs to learn)
2) All of your cucumber steps are specific to the one scenario you are testing - and now to modify any scenario you have to change the cucumber scenario, and then track down the definition and update that

The upside:

If your devs do it right you have a good suite of full-stack integration tests. Which is invaluable.
PREVIOUS NEXT REPLY QUOTE
 
Caltrops, please share your experiences with BDD or ATDD by Fullofkittens 04/17/2013, 4:37pm PDT NEW
    My experience: none by Entropy Stew 04/17/2013, 9:15pm PDT NEW
        Pretend that "doesn't" wasn't there NT by Entropy Stew 04/17/2013, 9:16pm PDT NEW
    Re: Caltrops, please share your experiences with BDD or ATDD by A MYSTERY SHOPPER 04/17/2013, 9:17pm PDT NEW
        This doesn't match my concept of ATDD or how I'm trying to apply it by Fullofkittens 04/18/2013, 4:25pm PDT NEW
    Re: Caltrops, please share your experiences with BDD or ATDD by E. L. Koba 04/17/2013, 9:45pm PDT NEW
        Re: Caltrops, please share your experiences with BDD or ATDD by A MYSTERY SHOPPER 04/17/2013, 9:54pm PDT NEW
            Re: Caltrops, please share your experiences with BDD or ATDD by E. L. Koba 04/18/2013, 1:50pm PDT NEW
            The issue is that we don't have enough automators to go around by Fullofkittens 04/18/2013, 4:32pm PDT NEW
        My knee-jerk reaction: JUSTIFIED by Entropy Stew 04/17/2013, 11:11pm PDT NEW
            Re: My knee-jerk reaction: JUSTIFIED by E. L. Koba 04/18/2013, 1:54pm PDT NEW
                Re: My knee-jerk reaction: JUSTIFIED by Fullofkittens 04/18/2013, 4:35pm PDT NEW
    Re: Caltrops, please share your experiences with BDD or ATDD by motherfuckerfoodeater 04/18/2013, 10:03am PDT NEW
        Re: Caltrops, please share your experiences with BDD or ATDD by Fullofkittens 04/18/2013, 5:33pm PDT NEW
 
powered by pointy