Package nz.ac.waikato.modeljunit.examples

Examples of Using ModelJUnit

See:
Description

Class Summary

AlarmClock
FSM	Simple example of a finite state machine (FSM) for testing.
LargeSet	A model of a set with N elements.
QuiDonc	A simple EFSM model of the Qui-Donc service.
SimpleSet	A model of a set with two elements: s1 and s2.
SimpleSetWithAdaptor	A model of a set with two elements: s1 and s2.
SizeExperiments	The performs some simple experiments on various test generation algorithms to compare the average size of the test suites that they must generate to achieve 100% transition coverage.
SmartSetAdaptor	This class shows how we can use inheritance to add adaptor code.
SpecialFSM	Simple FSM that is specialized so All Round Trips runs faster than Greedy or Random.
SpecialFSMNoLoops	Simple FSM that is specialized so All Round Trips runs better
StringSet	A simple implementation of a set of strings.
StringSetBuggy	A simple implementation of a set of strings.
StringSetTest	A simple example of a JUnit test that uses model-based test generation, from the model SimpleSetWithAdaptor.
TrafficLight

Enum Summary

QuiDonc.State
TrafficLight.STATE

Package nz.ac.waikato.modeljunit.examples Description

Examples of Using ModelJUnit

This package contains several simple examples of ModelJUnit models and test adaptors (the test harness/driver code that links the model to the system under test).

• SimpleSet. This is a very abstract (simplified) model of a set of elements. To simplify the testing, this model assumes that two elements is enough to test the set operations. It is only a model (contains no adaptor code), but it could be used to generate offline tests.
• SimpleSetWithAdaptor. This is the same as SimpleSet, but with some adaptor code added so that it does online testing of a Set<String> implementation whenever you generate tests from it. Pass any Set<String> object into the constructor to test it.
• LargeSet. This is a more complex model than SimpleSet, because it models up to 13 independent members of the set. The model is parameterized so that you can create models for any size of set from 1 to 13 elements. However, with 10 elements, this model has 1024 states and 20480 transitions, which would take a long time to test exhaustively. See SmartSetAdaptor for a better approach.
• SmartSetAdaptor. This illustrates how you can use inheritance to add adaptor code. This class extends SimpleSet and adds the adaptor code that interacts with the SUT (online testing). It also illustrates how you can do abstraction within the adaptor code. The adaptor code maps the model variable s1 to a single string (as in SimpleSetWithAdaptor), but maps the model variable s2 to a large number of strings. This achieves the goal of testing the SUT with a large number of elements, but still keeps the model small (4 states and 16 transitions) so that it can be tested quickly and thoroughly.
• StringSet and StringSetBuggy. These are example implementations of Set<String>, which you can use to experiment with executing the tests generated from the above models. For example, if we pass a StringSet object to SimpleSetWithAdaptor it passes all the tests, while if we pass StringSetBuggy it fails the tests.
• FSM. This is a trivial nonsense model with just three states, one of which is a terminal state. A graphical view of this simple model can be seen on the ModelJUnit web page.
• QuiDonc. This is a simple model of the France Telecom service that lets you enter a phone number and find out the name and address of the person who owns that phone number. This example is discussed in detail in our book, "Practical Model-Based Testing: A Tools Approach".
• ECinema. The ecinema subdirectory contains a model of a simple web-based cinema ticket booking system.
• strong>GSM 11-11. The gsm subdirectory contains a model of a subset of the GSM 11-11 protocol that is used is communicate with the SIM chip within many mobile phones. There is also a sample implementation of a simulated SIM chip that can be used as the system under test. It is interesting to use this to measure the code coverage and mutation score of the generated test suites, as well as the model coverage.

marku@cs.waikato.ac.nz

Last modified: Fri Aug 31 15:21:46 NZST 2007