I want to admit that I don’t know how to manage testing
Submitted by Ainars Galvans on Wed, 26/11/2008 - 16:07.
general software testing
I barely understand how to manage testing that I do. But for years I have failed to understand how to make my team do job as efficient and effective as I do (or at least close to it). I have a luxury to not be afraid to be marked as incompetent or dilettante: my test manager reputation in my company is strong enough.
By manage I mean planning, organizing, staffing, leading or directing, and controlling. Scripted testing was a great way to fool myself into believing I can: plan (#of test cases), stuff (to execute all of tests), lead or direct (select which test cases to be executed today) and control (monitoring run-logs against test cases).
I red AST update June, 2008 in September. Something shifted in my mind since I’ve seen Geordie Keitt trying to articulate a myth about test management that “requires a library of test scripts for tasking and status reporting”.
Yes I know about the alternative way to manage: SBTM (session based test management). It is better! I feel that this is much better way to fool myself again. I will try to ground this statement in my next blogs, for now I think I will only say this – yes testing is not a standalone task on a software schedule. If you manage testing better you could get it done faster, but not because you do the same job faster, but because you allow developers to fix bugs in the right order, with less stress, spending less time investigating and as a result making less bugs so you have less bugs to find and less job as a result. If you manage it right you could end up with less regression tests cycles, less time wasted into reworking test documentation (including bug-reporting).
I will try to emphasize need to trust in self-managing testers in future blogs.
By manage I mean planning, organizing, staffing, leading or directing, and controlling. Scripted testing was a great way to fool myself into believing I can: plan (#of test cases), stuff (to execute all of tests), lead or direct (select which test cases to be executed today) and control (monitoring run-logs against test cases).
I red AST update June, 2008 in September. Something shifted in my mind since I’ve seen Geordie Keitt trying to articulate a myth about test management that “requires a library of test scripts for tasking and status reporting”.
Yes I know about the alternative way to manage: SBTM (session based test management). It is better! I feel that this is much better way to fool myself again. I will try to ground this statement in my next blogs, for now I think I will only say this – yes testing is not a standalone task on a software schedule. If you manage testing better you could get it done faster, but not because you do the same job faster, but because you allow developers to fix bugs in the right order, with less stress, spending less time investigating and as a result making less bugs so you have less bugs to find and less job as a result. If you manage it right you could end up with less regression tests cycles, less time wasted into reworking test documentation (including bug-reporting).
I will try to emphasize need to trust in self-managing testers in future blogs.