Software Engineering Software Testing Strategies867

7/30/2019 Software Engineering Software Testing Strategies867

http://slidepdf.com/reader/full/software-engineering-software-testing-strategies867 1/35

Software Engineering – Software Testing Strategies

STRATEGIC APPROACH

Testing is a set of activities that can be planned in advance and conducted

systematically.

Many strategies exist, all follow these generic characteristics:

To perform testing, a software team should conduct effective formal technicalreviews. This eliminates errors before testing commences.

Testing begins at the component level and works outward toward the integration of the entire computer-based system.

Different testing techniques are appropriate at different points of time.

Testing is conducted by the developer of the software and (for large projects) anindependent test group.

Testing and debugging are different activities, but debugging must beaccommodated in any testing strategy.




VERIFICATION VS. VALIDATION

Verification refers to the set of activities the ensures that software correctly

implements a specific function.

Validation refers to a different set of activities that ensures that the software that hasbeen built is traceable to customer requirements.

Verification: Are we building the product right?

Validation: Are we building the right product?




ORGANIZING TESTING

“Optimism is the occupational hazard of programming; testing is the treatment.”

Kent Beck

“The first mistake that people make is thinking that the testing team is responsible forassuring quality”




ORGANIZING TESTING

Software testing may be viewed as a spiral process:




ORGANIZING TESTING

Software Testing Steps




WHEN IS TESTING COMPLETE?

How do you know when you are done testing?

There is no definitive answer to this question.

Many possible answers:

You are never done testing.

You are done testing when you run out of money or time.

Depends upon the certainty of failure you desire.




STRATEGIC ISSUES

The following issues must be addressed if a successful software testing strategy is to

be implemented:

Specify product requirements in a quantifiable manner long before testingcommences.

State testing objectives explicitly.

– test coverage, mean time to failure, cost to find and fix defects, remaining defect density or frequencyof occurrence

Understand the users of the software and develop a profile for each usercategory.

Build robust software that is designed to test itself.

Use effective formal technical reviews prior to testing.

Conduct formal technical reviews to assess the test strategy and test casesthemselves.

Develo a continuous im rovement a roach for the testin rocess.




TEST STRATEGIES

Unit Testing – focuses verification effort on the smallest unit of software design. (The

component or module)

Considerations:

The module interface is tested to ensure that information properly flows intoand out of the program unit under test.

Local data structures are examined to ensure that data stored temporarilymaintains its integrity during all steps in an algorithm’s execution.

All independent paths through the control structure are exercised to ensure thatall statements in a module have been executed at least once.

Boundary conditions are tested to ensure that the module operates properly.

Error handling paths are tested.




TEST STRATEGIES

Start testing data flow across the module interface.

Exercise all local data structures, test impact on global structures.

Common errors include:

misunderstood or incorrect arithmetic precedence

mixed mode operations

incorrect initialization

precision inaccuracy

incorrect symbolic representation of an expression




TEST STRATEGIES

UNIT TESTING PROCEDURES

Unit testing is normally considered as an adjunct to the coding step.

Design of tests can be done before coding begins (Agile programming)




TEST STRATEGIES





TEST STRATEGIES


Since each component is not a stand alone program, a bit of work must be done toperform the test.

Create a driver, usually a main program, that accepts test case data and passes it otthe component to be tested. It also may print relevant results.

Stubs replace modules that may not be coded and are subordinate to the componentto be tested. A stub may do minimum data manipulation , provides verificationof entry, and returns control to the module undergoing testing.




TEST STRATEGIES


Unit testing is simpler when modules are highly cohesive.




TEST STRATEGIES

INTERGRATION TESTING

Why is unit testing not enough?

Data can be lost across an interface.

One module can have an inadvertent effect on another.

Global data structures can present problems

Individually acceptable imprecision may become magnified.

Resource issues.




TEST STRATEGIES


Integration testing is a systematic technique for constructing the software architecturewhile at the same time conducting tests to uncover errors associated withinterfacing.

The objective is to take unit tested components and build a program structure that has

been dictated by design.




TEST STRATEGIES


There is a tendency to attempt nonincremental integration – construct the wholeprogram via a “big-bang” approach.

Thus to combine all components at once and test it as an entire program.

S f S f S




TEST STRATEGIES


Top Down Integration is an incremental approach to construction of the softwarearchitecture. Modules are integrated by moving downward through the controlhierarchy, beginning with the main control module.

Modules subordinate to the main control module are incorporated into the structure in

either a depth first or breath first manner.

S ft E i i S ft T ti St t i




TEST STRATEGIES - INTERGRATION TESTING

The integration process is preformed in five steps:

1. The main control module is used as a test driver, and stubs are substituted forall components directly subordinate to the main control module.

2. Depending on the integration approach selected (i.e. depth or breath first),subordinate stubs are replaced one at a time with actual components.

3. Tests are conducted as each component is integrated.

4. On completion of each set of tests, another stub is replaced with the realcomponent.

5. Regression testing may be conducted to ensure that new errors have not beenintroduced.






Verifies major control or decision points early

Problems exist however:

When processing at low levels in the hierarchy is required to test upper levels stubsreplace low-levels in the hierarchy, therefore no significant data can flowupward in the program structure.

Three options:

1. delay many tests until stubs are replaced with actual models

2. develop stubs that perform limited functions that simulate the actual module

3. integrate the software from the bottom of the hierarchy upward

1. The first method causes us to lose control over specific tests and specificmodules. Thus leading to difficulty determining the cause of errors and violatesthe nature of the top down approach.

2. The second method is workable, but leads to significant overhead.

3. The third approach is called bottom-up testing, to be covered shortly.






Bottom Up Integration

Begins construction and testing with atomic modules. Since all subordinates arealways available, there is no need for stubs.

Bottom up integration follows the following steps:

1. Low-level components are combined into clusters that perform a specificsoftware subfunction.

2. A driver (control program for testing) is written to coordinate test cases inputand output.

3. The cluster is tested.

4. Drivers are removed and clusters are combined moving upward in the program

structure.

One major disadvantage of bottom up integration is that the program as an entitydoes not exist until the last module is added.






Bottom Up Integration






Regression Testing

Each time a new module is added, the software changes.

These changes may cause problems with functions that previously worked.

Therefore, we must re-execute some subset of tests that were already conducted to

ensure that changes have not propagated unintended side effects.

Any time you test, if successful you will find errors. When those errors are corrected,software changes. Therefore, regression testing must be performed in order toensure that those changes did not propagate unintended side effects.

Regression tests may be conducted manually – I do not recommend this.






Integration Test Documentation

The overall plan for integration of the software and a description of specific tests aredocumented in a Test Specification.

The document contains:

a test plan

a test procedure

is a work product of the software process

becomes a part of the software configuration

A test plan describes the overall strategy of integration.

Testing is divided into phases and builds that address specific functional andbehavioral characteristics of the software.

Soft are Engineering Soft are Testing Strategies




TEST STRATEGIES FOR OBJECT-ORIENTED SOFTWARE

Unit Testing in the OO Context

With OO the concept of the unit changes.

Because of the dependence of sub classes, unit testing is a bit more complicated.

In the OO context, unit testing is basically Class testing.

Software Engineering Software Testing Strategies





Integration Testing in the OO Context

Unlike non OO programming, OO programming does not have an hierarchical controlstructure.

Integrating one object at a time is also difficult because of the direct and indirectinteractions of the components that make up the class.

There are two basic strategies:

Thread-based

Use-based






Thread-based testing integrates the set of classes required to respond to one input orevent for the system.

Each thread is integrated and tested individually.

Regression testing is applied to ensure no side effects occur.

Use-based testing begins the construction of system by testing those classes that usevery few server classes.

After independent classes are tested, the next layer of classes, called dependentclasses are tested.

This continues until the entire system is constructed.






Drivers can be used to test operations at the lowest level and for testing groups of classes.

A driver may be used to replace the user interface.

Stubs may be used where collaboration between classes is required but one or more of the collaborating classes is not fully implemented.






Cluster testing is one step in integration testing of OO software.

Here, a cluster of collaborating classes is exercised by designing test cases thatattempt to uncover errors in the collaborations.






Validation Testing

Validation testing begins at the culmination of integrations testing.

Testing focuses on the user-visible actions and user-recognizable output from thesystem.

Validation can be defined as:

Validation succeeds when software functions in a manner that can bereasonably expected by the customer.

Software validation is achieved through a series of tests that demonstrate conformitywith software requirements.






Most software product builders use a process called alpha and beta testing to uncovererrors that only the end-user seems able to find.

The alpha test is conducted at the developer’s site by end-users.

The beta test is conducted at the end-user sites. Beta tests usually do not have thedeveloper present.






System Testing

Recovery Testing – most computer-based systems must recover from faults andresume processing within a prespecified time.

Recovery testing is a system test that forces the software to fail in a variety of waysand verifies that recovery is properly preformed.






Security Testing –

Any computer system that manages sensitive information or causes actions that canimproperly harm or benefit individuals is a target for improper or illegalpenetration.

Lot’s of security issues.

Security testing verifies that protection mechanisms built into the system will, in fact,protect it from improper penetration.

During security testing, the tester plays the role(s) of individual who desires topenetrate the system.

Anything goes!

What is the biggest source of security issues?






Usually it’s not the software, but people issues.

People often give out secure information. No system can truly protect against this100%.






Stress Testing

So far all testing assumed normal program functions and performance.

Stress tests are designed to confront programs with abnormal quantity, frequency, orvolume.

Multiple interrupts per second Increased data rate

Test max memory

Excessive disk hunting






Performance Testing

Designed to test the run-time performance of software within the context of anintegrated system.

Should be performed at all steps of the testing process.

Not until system elements are fully integrated that performance of a system can reallybe ascertained.

Sometime coupled with stress testing.

Often requires hardware and software instrumentation to measure resource utilization.

Documents

Software Engineering Software Testing Strategies867