robust test design - a mindmap
DESCRIPTION
This mind map was presented at HBT Master Class, Aug 9, 2012.TRANSCRIPT
Robust Test DesignPowered by HBT
ChallengesAre my test casesadequate?
Do I have the leastnumber of test cases?
Fishinganalogy
Objective: Catch fishesthat matter, efficiently
HOW?Adopt a
SCIENTIFICAPPROACH
What "fishes" to catch?i.e PDT to look for
Where to "fish"?i.e What Req, Feature,
Component?
What kind of net?i.e. What test techniques?
How to cover quickly?What tools and what TS?
Testing-Scientific Approach
Concepts
Cleanliness criteriaAre the expectations ofend-users
"Business goal"
Potential Defect Types(PDT)
Classes of defectsthat may be probable
"Technical goal"
Quality levelsNOT Test levels
Test levels
Unittesting
Integrationtesting
Systemtesting
Qualitylevels
Clear notion of"HOW CLEAN"
NINE levels
Each level focusedon uncoveredcertain PDTs
Test types
Access controltest
Input validationtest
Interfacevalidation test
Interaction test
Structure testGood citizen
test
Functionalitytest
Load, Stresstest
Installation testPerformance
test
Compatilibitytest
Scalability test
Migration test
Volume test
Usability test
Security testEnd-to-End
operations test
Test techniques
Enables design ofTS/TC
Designtechniques
Classification based onPROPERTY
"EXTERNAL"Behaviour based
"INTERNAL"Structure based
Classification based onOUTCOME
Scenario generationData generationTest case generationSTEM Core Concept
"Techniques landscape"
Test scenario
Is a flow
Represents a behaviour
Behaviour emanatingfrom a combination of
conditionsPositive - Correct use
scenario
Negative - Abusescenario
Scenario is what isautomated
Is a collection of testcases
#TS governed by#conditions
Test case
Is a "stimulus"
To excite theSUT
... to invoke abehaviour
Positive
All data valuesvalid
Is always partof +ve TS
Negative
At least ondata value
invalid
Is always partof -ve TS
AdequacyRequirements traceabilityFault traceability Code coverage
QL-CC-TTmapping
Documentation
ObjectivePre-requisites
Execution stepsData set(s) to be used
Expected outcome
Properties ofgood TS/TC
Designprocess
1. Choose quality level.
2. Identify conditionspertaining to the level.
Descriptiveto
Prescriptive3. Generate scenarios.4. For each scenario,generate data valuesfor each input.5. Combine the dataoptimally to generatetest cases.
6. Trace thescenarios/cases toPDTs and ensure thatthe PDTs pertaining tothat level areuncoverable.