Testing as analysis is when we find new things using testing techniques. The goal of this activity is analysis, not serendipitous finding of functionality during validation or verification. This kind of testing is performed during GorillaTesting or AdHocTesting.

In the role of tester, the AvailableCustomer often discovers new requirements. 
	* Things do not work as they should, but do meet existing requirements. 
	* holes in requirements
	* extensions that were not thought of previously
	* new stuff that just shows up during testing

This is the idea that something can be DoneDoneDone but still not be right or meet existing requirements. AdHocTesting is not validation; it is analysis.

-- DanRawsthorne, DouglasShimp

Thoughts?

----
''In AgileDevelopment, this is basically the feedback loop from frequent deliveries.''

-- Author Unknown?
----

There are many feedback loops in AgileDevelopment. TestingAsAnalysis is a useful distinction in handling some organizational impediments.

'''Why?'''

Often we see testers doing their job and logging defects against the code base when, in fact, these are not defects but new behavior that is wanted. We might have tried during analysis (e.g. writing use cases, stories, feature decompose etc.) to identify all of the important behaviors that we wanted to handle but are not found until testing. Testers are not recognized for their analysis work. As a result, a common set of impediments to agility occurs. 
For example, 
	* Testers log a defect and developers argue that it is a new requirement but, the testers maintain that it is a defect. 
	* Many large organizations are interested in adopting an agile process but, for them to move towards agile their requires explanations of sufficient detail that explains this problem well. 
	* Another risks in alienating testers from the WholeTeam and loosing this value of HighBandwidthCommunication. 
If this problem is not handled well, the entire organizational situation is prone to ConflictBlamePolitics.


-- DouglasShimp
----
Let me echo what Doug says. In a normal (bad) project these defects cause scope creep ''within'' a story, which is forbidden - stories have fixed scope. By saying that these are automatically new stories, and that the testers are doing TestingAsAnalysis, we have produced a new paradigm for the testers. While this is something that we (as agilists) have known for years, it is new for them...

-- DanRawsthorne

----
CategoryTesting
CategoryAnalysis