I'd be interested to hear views on how people test web based applications, specifically when the client is reluctant to allocate suitable time frames for testing.
I've been looking into Rapid and Exploratory testing (specifically James Bach's research) and feel that implementing these methodologies would benefit some of our customers substantially. However, for peace of mind and accountability we would need to ensure test coverage to a certain extent.
Can anyone provide any tips for implementing Rapid/exploratory testing or suggest another method?
When I was last challenged with this, I created a checklist of the test names and used that as an inventory. I combined that with normal ad hoc and/or exploratory testing. If there was to be additional downstream testing, I would feed that team with coverage information for them to leverage.
Equally important was the need to have the customer sign off on the risks and inventory of tests-to-be-executed. Within the context of your dilemma, I always make it a point for the risk-taker to share in the risk. When I reported defects, I did that on the spot and made it a point to add more repeatability detail. It worked well for me.
Originally posted by rosiefrancis: I'd be interested to hear views on how people test web based applications, specifically when the client is reluctant to allocate suitable time frames for testing.
<font size="2" face="Verdana, Arial, Helvetica">If you rank your tests in an order that would be meaningful to your client (perhaps in order of decreasing risk), and estimate the time required for each test, then you and your client can jointly draw the line at the point where you should stop testing.
That way, you'll know what to test, and your client will know what you won't be testing.
If you're in all THAT much of a hurry, you might not have time to read it, but I would recommend Rick Craig and Stefan Jaskiel's book "Systematic Software Testing." It addresses risk-based testing in some detail.
N Rao, I don't understand how that addresses the original question (which is 2 months old, by the way). She is not asking what to test, but what approach to use when insufficient time has been allowed for testing.