Download - CSP 587 Week 5 - Testing
-
8/10/2019 CSP 587 Week 5 - Testing
1/15
Transform ing Lives. Invent ing the Future.
www.i i t .edu
I ELLI NO IS T UI NS TI T
O F TEC H N O LO G Y
SOFTWAREQUALITYMANAGEMENT
CSP 587
Dennis Hood
Fall 14
-
8/10/2019 CSP 587 Week 5 - Testing
2/15
Week 5Testing
-
8/10/2019 CSP 587 Week 5 - Testing
3/15
Checks and Balances
Work products should be reviewed to discover defects
Work products come in many forms
So do defects
Undiscovered defects will be passed downstream
Defect discovery is much more effective when the
reviewer is independent of the production team
Why?
How do we achieve necessary independence while
maintaining sufficient knowledge?
-
8/10/2019 CSP 587 Week 5 - Testing
4/15
Case Study: Challenger
-
8/10/2019 CSP 587 Week 5 - Testing
5/15
Case Study: Challenger
-
8/10/2019 CSP 587 Week 5 - Testing
6/15
-
8/10/2019 CSP 587 Week 5 - Testing
7/15
-
8/10/2019 CSP 587 Week 5 - Testing
8/15
Review Objectives
Direct objectivesrelated to the current project
Detect defects, including deviations from standards
Identify new risks
Obtain approval to move forward
Indirect objectiveslong-term team growth
Facilitate the exchange of professional knowledge regarding
development
Data collection for future process improvement
-
8/10/2019 CSP 587 Week 5 - Testing
9/15
Formal Design Reviews
Conducted at milestones requiring the completion of
analysis or design products
Participants
Independent, senior leader
Review team (primarily senior personnel)
Approval decisions
Full, partial, or denial Document the outcome
Including responsibilities for required corrections
Follow up
-
8/10/2019 CSP 587 Week 5 - Testing
10/15
Peer Reviews
Inspections and walkthroughs
Less formal than formal reviews
Not as much independence Not as much senior authority
More focused on defect detection than approval
Not as likely to serve as a milestone gate
Easier to organize, not as bureaucratic, etc.
-
8/10/2019 CSP 587 Week 5 - Testing
11/15
Expert Opinions
Significantly independent
Either external review or participation in
internal reviews Beneficial when:
In-house expertise is absent or unavailable
Internal deadlock due to disagreement by seniorpersonnel on significant issues
-
8/10/2019 CSP 587 Week 5 - Testing
12/15
Software Testing
Traditionally applied late in the SDLC
Common sense has moved it up
Testing differs from reviews in that it involves
exercising work products Specialized testing team
Approved test procedures
Approved test cases
Objectives Direct
Discover defects and achieve acceptability
Indirect Collect data
-
8/10/2019 CSP 587 Week 5 - Testing
13/15
Strategies
Big-bang vs. incremental
More comprehensive and final testing of the full
package, vs.
Sooner, quicker yet less complete testing of
modules
Top-down vs. bottom-up
Start at the main module and drive down into
the functionality, vs.
Focusing on the details first
-
8/10/2019 CSP 587 Week 5 - Testing
14/15
Classifications
Black-box
Focus on output
Equivalence classes (sets of inputs producing the same output)
Boundary values
Documentation testing
Availability and reliability
Stress
Security
Usability
White-box
Focus on paths through the code
Correctness of calculations Path and line coverage
Cyclomatic complexity metrics
Qualification and reusability
-
8/10/2019 CSP 587 Week 5 - Testing
15/15
Implementation
The testing process
Determine the test methodology
Plan the tests
Design the tests
Test away!
Test cases
Automation
Alpha and beta testing