On Dec 28, 2011, at 11:02 AM, Joel Foner <joel.foner@gmail.com> wrote:
> How would you define a failed test automation project? One that only achieved partial coverage? One that got outright cancelled?
I thought of one we'd all call a failure. A guy who no one could communicate with was hired to write tests to cover a relatively small existing system. He left before that was done and the existing developers (2 coders + a manual tester) just stubbed out most of his tests because they couldn't make any sense out of them. A few months later the remaining few finally agreed they should have been writing tests themselves.
Back in my aerospace days automated test was absolutely essential but it was routinely added later in development than anyone would advocate today. Though far from optimal, that effort added value and helps keep you safe. I suppose it was possible to get away with this back then because the requirements were much more stable and huge amounts of effort (several times the design/code effort) also went into reviewing and manual testing. So, pretty wasteful but not a total failure.
--mj
Sent from a phone that often corrects words I tapped to words I may not have meant.
0 comments:
Post a Comment