Analysis/Testing

Analysis/Testing

1. V-Model

   • Coding
     • Program Specification
       • Technical Specification
         • Functional Specification
           • Requirement Specification
     • Unit Testing
       • Integration Testing
         • System Testing
           • Acceptance Testing
     • Unit Test Planning
       • Integration Test Planning
         • System Test Planning
           • Acceptance Test Planning
   • Testing starts at the beginning
   • Advantages
     • Faults prevented
     • Improved risk management
   • Disadvantages
     • Rigid - not agile
     • No early prototypes

2. T-Tests

   • Variance
     • Take mean from each value in set
     • Square each
     • Add together
       • Sample
         • Divide by number of values -1
         • Σ x^2/n-1 - Σ (x)^2/n(n-1)
       • Population
         • Divide by number of values
         • Σ x^2/n - Σ (x)^2/n(n)
   • Standard Deviation
     • Square root of variance
   • Standard Error
     • Std Dev / √n
     • Std Dev / √n-1
   • Independent
     • 1 tailed
       • Effect in certain direction
   • Paired
     • Data from one sample linked with another
     • 2 tailed
       • Effect different
   • Steps
     • Define h0 and h1
     • State Alpha
       • Accuracy
     • Calculate DOF
       • num of values -1
     • State decision rule
       • 1 or 2 tails
     • Calculate test statistic
       • (sample mean - pop mean) / standard error
     • State results
     • State conclusion
       • If t stat > t value, reject h0
         • Data lies within critical region
       • t(15) = 3.1, p < 0.005
         • t(DOF) = T-STAT, p < ALPHA
   • Confidence Intervals
     • mean ± (sample std dev / √n) * t statistic * (n-1)

3. Black Box

   • Equivalence Partitioning
     • Divide inputs into blocks
     • Values in each block are equivalent
     • Each value can only be in 1 block
     • All values must be represented
     • Create as few blocks as possible
     • Equivalence Class Partitioning
       • Valid class(es)
       • Invalid class(es)
       • Partition into disjoint subsets
   • Boundary Value Analysis
     • Boundary -1
     • Boundary value
       • Of each edge
     • Boundary +1
     • Robustness Testing
       • Negative testing - Cases where boundaries are slightly exeeded
       • Worst case
         • 9 additional cases at each overlap
   • Functionality testing
     • Pre-condition
     • Post-condition
   • Combinational Design
     • Subset of combinations generated
     • Reduces number of test cases
     • Each choice coverage
       • Each characteristic must be in at least 1 test case
       • (A1,B1,C1) (A2,B2,C2) (A3,B3,C3)
     • Pair-wise testing
       • Number of test cases: (parameters ) ^ values
     • Special cases / Error Guessing
       • Use judgement to guess where errors are
     • State-based testing
       • For state-based systems
       • State model
         • States
         • Transitions
         • Events
         • Actions

4. White Box

   • Every path is identified and tested
   • Path Analysis
     • Control Flow Graph
       • Cyclomatic complexity (C)
         • Number of paths needed
         • Number of regions (enclosed areas, + whole gaph) OR number of predicates (nodes with multiple exit arcs) + 1
       • Path based testing
         • Number of test cases < C
         • Cover all statements
         • Exercise all decisions
     • Path
       • Collection of Nodes with Directed edges
   • Code Coverage
     • Decision/Branch
       • True case and false case
     • Condition
       • True/false cases for every outcome
     • Multiple Condition
       • True/false cases for every condition
     • Loop
     • Line
       • Every line of code tested
       • Simple, but few test cases

5. Scientific Method

   • Mill's methods
     • Agreement
       • Only 1 thing all cases with effect share is the cause
     • Disagreement
       • Only thing case with effect has is the cause
     • Joint
     • Concomitant Variation
       • Change of cause correlates with change of effect
     • Residues
       • Take away until remaining effect is caused my remaining circumstances
   • Steps
     • Question
     • Observation / Research
     • Hypothesis
       • Testable and falsifiable
     • Prediction
     • Test / Experiment
       • Test h0
     • Analysis
     • Communication

Imágenes y archivos del mapa