01881nam a2200181 a 450000100080000000500110000800800410001910000170006024501250007726002420020230000160044452011490046065300190160965300220162870000210165070000130167170000150168410085482020-01-17       2002    bl uuuu     u00u1 u    #d1 aCHAIM, M. L.  aData-flow testing of programs with pointers and recordsbimplementation, effectiveness, and cost.h[electronic resource]  aIn: INTERNATIONAL CONFERENCE ON COMPUTER SCIENCE, SOFTWARE ENGINEERING, INFORMATION TECHNOLOGY, E-BUSSINESS, AND APPLICATIONS, 2002, Foz do Iguazu. Proceedings. [S.l.]: International Association for Computer and Information Sciencec2002  ap. 317-323.  aA conservative data-flow model for implementation of accurate coverage assessment of pointers and records in data-flow resting is described. To assess its cost and effectiveness, we evaluated the adequacy of tests with respect to two different data-flow testing criteria implemented using the conservative data-flow model for seven faulty versions of the Unix sort program. The seeded faults simulate the misure os pointers and individual fields of records. The effectiveness and size of these adequate test sets were compared against adequate test sets determined using the same criteria implemented with a data-flow model that do not treat pointers and records in any special way. Results of the comparison indicate that the effectiveness of the conservative data-flow models depends on the program and the fault(s) it contains. Moreover, there exist faults whose detection is facilitated by a more accurate data-flow analysis such as that proposed by the conservative data-flow model. For one of the testing criteria investigated, we also found that the extra cost of analysis and coverage measurement for pointers and records is reasonable.  aFluxo de dados  aTeste de software1 aMALDONADO, J. C.1 aJINO, M.1 aVILELA, P.