E. Brinksma, J. Tretmans and L. Verhaard, A Framework for Test Selection, Proc. PSTV, pp. 233248, 1991..  Home/Search   Document Not in Database   Summary   Related Articles   Check

....that a particular relation is a symmetry on the system, but for the black box implementation one has to assume that this is the case. The reliability of this assumption is the tester s responsibility. In this respect, one may think of exploiting symmetry as a structured way of test case selection [13, 4] for systems too large to be tested exhaustively, where at least some subautomata are tested thoroughly. This paper is not the first to deal with symmetry in protocol testing. In [22] similar techniques have been developed for a test generation methodology based on labeled transition systems, ....

E. Brinksma, J. Tretmans, and L. Verhaard. A framework for test selection. In B. Jonsson, J. Parrow, and B. Pehrson, editors, Protocol Specification Testing and Verification, Volume XI, pages 233--248. North-Holland, 1991.

....space of a specification, but how much it covers the whole behaviour of the specification is unknown. Moreover, when the circuits are relatively complex, the size of the test suite might be very large resulting in considerable simulation run time. The methodologies of test selection advocated in [BTV91, ACV93, CG97] might help. Heuristics related to circuits should also be explored, which is extremely useful for testing important parts of a circuit, or the parts that are subject to errors. 92 ....

Ed Brinksma, Jan Tretmans, and Louis Verhaard. A framework for test selection. In Bengt Jonsson, Joachim Parrow, and Bjrn Pehrson, editors, Proc. Protocol Specification, Testing and Verification XI, pages 233--248, Amsterdam, Netherlands, 1991. North-Holland.

.... Interaction Problem ( i; flashhookcw(A) flashhooktwc(A) exit [ i; hangup (A) exit [ i; flashhooktwc(A) flashhookcw(A) exit ) The final result is shown in Figure 5 (b) The next step is to express T s as a set of irreducible test cases, from which a set of useful test cases are selected[BrTV91]. Some such test cases are: T 1 = answer(B, A) flashhookcw(A) flashhooktwc(A) exit T 2 = answer(B, A) hangup (A) exit T 3 = answer(B, A) flashhooktwc(A) flashhookcw(A) exit T 4 = answer(B, A) flashhookcw(A) flashhooktwc(A) exit [ hangup (A) exit [ flashhooktwc(A) ....

....in this very simplified example we have assumed finite behavior trees, the presence of loops is not a problem [Jao92] DAV93] 3. 4 Executing the System and Analysing the Results (step and ) The final two steps of the methodology are to execute the specification against a selected subset [BrTV91] of the derived test suite in order to check for deadlocks. For our purposes, the set T 1 , T 2 , T 3 is sufficient to test our integration because every trace of T 4 is a member of another test suite in the selected set. An example of testing the integration with T 1 is shown below. Testing ....

[Article contains additional citation context not shown here]

E. Brinksma, J. Tretmans, L. Verhaard. A Framework for Test Selection, Protocol Specification, Testing, and Verification, XI. Eds. B. Jonsson, J. Parrow, and B. Pehrson, North Holland 1991, 233-248.

....that a particular relation is a symmetry on the system, but for the black box implementation one has to assume that this is the case. The reliability of this assumption is the tester s responsibility. In this respect, one may think of exploiting symmetry as a structured way of test case selection [13, 4] for systems too large to be tested exhaustively, where at least some subautomata are tested thoroughly. This paper is not the first to deal with symmetry in protocol testing. In [22] similar techniques have been developed for a test generation methodology based on labeled transition systems, ....

E. Brinksma, J. Tretmans, and L. Verhaard. A framework for test selection. In B. Jonsson, J. Parrow, and B. Pehrson, editors, Protocol Specification Testing and Verification, Volume XI, pages 233--248. North-Holland, 1991.

....inputs to the machine and observing the outputs produced. There is an extensive literature on testing finite state machines, starting from Moore s seminal paper on gedanken experiments [24] motivated first from circuits (see [11, 16] and more recently from testing of communication protocols [2, 3, 4, 5, 9, 18, 19, 21, 22, 28, 29, 31, 32, 33]; see [20] for a survey and more complete bibliography. For the testing problem considered in this paper, the machine M is completely specified except for its initial state. We are given a possible set of choices for the initial state, and the goal is to design an experiment that provides inputs ....

E. Brinksma, J. Tretmans, and L. Verhaard, A framework for test selection. In Proc. IFIP WG6.1 11th Intl. Symp. on Protocol Specification, Testing, and Verification, pp. 233-48, 1991.

....that a particular relation is a symmetry on the system, but for the black box implementation one has to assume that this is the case. The reliability of this assumption is the tester s responsibility. In this respect, one may think of exploiting symmetry as a structured way of test case selection [13, 4] for systems too large to be tested exhaustively, where at least some subautomata are tested thoroughly. This paper is not the first to deal with symmetry in protocol testing. In [20] similar techniques have been developed for a test generation methodology based on labeled transition systems, ....

E. Brinksma, J. Tretmans and L. Verhaard. A framework for test selection. In B. Jonsson, J. Parrow and B. Pehrson, editors, Protocol Specification Testing and Verification, Volume XI, pages 233-248. North-Holland, 1991.

....jModO (P )j 6= Our definition is based on the intuition that only the interpretation of observable sentences by an oracle framework can give semantics to programs. Our definition of correctness is similar to the one of implementation in the area of conformance testing for protocol verification [6, 16] (they say P implements SP ) Remark. The program Q 2 (Example 2) is a typical example where the program under test exports less predicates than the specification. Lists may not be observable when for example, lists are implemented with pointers. On the other hand, if lists are a predefined ....

E. Brinksma, J. Tretmans, and L. Verhaard. A framework for test selection. In Procol Specification, Testing, and Verification XI, North Holland, 1991.

.... distributed systems design were first presented in [BAL 90, Bri91] Other frameworks, for the formalization of software testing, can be found in [BGM91, Gou83, How78] Work on informal coverage measures can be found in many places, e.g. RW87] A metric treatment of coverage was proposed in [BTV91] with subsequent work in [Bri93] Another metric approach to coverage is presented in [ACV93] 12. Milestones and Deliverables: a) contributions to the ISO ITU Standard Recommendation on Formal Methods in Conformance Testing (b) regular publications in leading conference proceedings and ....

Ed Brinksma, Jan Tretmans, and L. Verhaard. A framework for test selection. In B. Jonsson, J. Parrow, and B. Pehrson, editors, Protocol Specification, Testing, and Verification, XI, pages 233--248. IFIP WG 6.1, North-Holland Publishing Company, 1991. Also: Memoranda Informatica 91--54, TIOS 91/022, University of Twente.

....that a particular relation is a symmetry on the system, but for the black box implementation one has to assume that this is the case. The reliability of this assumption is the tester s responsibility. In this respect, one may think of exploiting symmetry as a structured way of test case selection [13, 4] for systems too large to be tested exhaustively, where at least some subautomata are tested thoroughly. This paper is not the first to deal with symmetry in protocol testing. In [20] similar techniques have been developed for a test generation methodology based on labeled transition systems, ....

E. Brinksma, J. Tretmans and L. Verhaard. A framework for test selection. In B. Jonsson, J. Parrow and B. Pehrson, editors, Protocol Specification Testing and Verification, Volume XI, pages 233-248. North-Holland, 1991.

No context found.

E. Brinksma, J. Tretmans and L. Verhaard, A Framework for Test Selection, Proc. PSTV, pp. 233248, 1991..

....limit than by using a larger label subset. But this is not always true, e.g. increasing l c from 101 to 102 might give lesser increase than taking a larger labelset. Moreover we de ned speci c values for p k and dL . It can be seen that in this example the monotonicity property required in [4] viz. T T 0 ) cov(T ) cov(T 0 ) is respected. From an intuitive point of view this property is reasonable: if one wants a better coverage, one needs to generate more tests. This property can also be proven to hold in general. 8. Conclusions A heuristic is a general guideline for reducing ....

E. Brinksma, J. Tretmans, and L. Verhaard. A framework for test selection. In B. Jonsson, J. Parrow, and B. Pehrson, editors, Protocol, Specication, Testing, and Verication, volume XI, pages 233-248. North-Holland, 1991.

....made by implementers, which kind of errors are important, e.g. in the sense of having catastrophic consequences, what functionality is dicult to implement, which functionality is crucial for the well functioning of the system, etc. An approach to formalizing this extra information was given in [BTV91] On the other hand, assumptions can be made about the occurrence of errors in implementations, e.g. that errors will not occur in isolation, i.e. if some behaviour is erroneous then there is a large probability that some other behaviour close to it is also erroneous. So we only have to test ....

....it is better to increase the cycling limit for obtaining a better coverage. But this is not always true because we de ned speci c values for p k and dL (for other values, to increase the label subset will be better) It can be seen from this example that the monotonicity property required in [BTV91] that T T 0 ) cov(T ) cov(T 0 ) is respected by our coverage. From an intuitive point of view this property is reasonable: if one wants a better coverage, one needs to generate more tests. We want to prove it in the general case. For this we will make an assumption which is quite ....

E. Brinksma, J. Tretmans, and L. Verhaard. A framework for test selection. In B. Jonsson, J. Parrow, and B. Pehrson, editors, Protocol, Specication, Testing, and Verication, volume XI, pages 233-248. North-Holland, 1991. 18

....will only continue up to a maximal property N for which P ffl N can be established in a reasonable time if it holds, and the residual risk of inconclusiveness of W(C; Gamma [ N ] is acceptable. Determining such optimal N can be done following similar strategies as discussed in [Tre92, BTV91], but we will not discuss those here. Instead we will study how approximating sequences i may be constructed. The situation for the test scenario and or falsification approaches is, mutatis mutandis, the same. 3 Instantiation of the validation framework 3.1 Modelling processes and ....

Ed Brinksma, Jan Tretmans, and Louis Verhaard. A framework for test selection. In B. Jonsson, J. Parrow, and B. Pehrson, editors, Protocol Specification, Testing, and Verification, XI, pages 233--248. North-Holland Publishing Company, 1991.

No context found.

E. Brinksma, J. Tretmans, and L. Verhaard, "A framework for test selection," Proc. IFIP WG6.1 11th Intl. Symp. on Protocol Specification, Testing, and Verification, North-Holland, B. Jonsson, J. Parrow, and B. Pehrson Ed. pp. 233-248, 1991.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC