简介:Debuggingisatime-consumingtaskinsoftwaredevelopment.Althoughvariousautomatedapproacheshavebeenproposed,theyarenoteffectiveenough.Ontheotherhand,inmanualdebugging,developershavedifficultyinchoosingbreakpoints.Toaddresstheseproblemsandhelpdeveloperslocatefaultseffectively,weproposeaninteractivefault-localizationframework,combiningthebenefitsofautomatedapproachesandmanualdebugging.Beforethefaultisfound,thisframeworkcontinuouslyrecommendscheckingpointsbasedonstatements'suspicions,whicharecalculatedaccordingtotheexecutioninformationoftestcasesandthefeedbackinformationfromthedeveloperatearliercheckingpoints.Thenweproposeanaiveapproach,whichisaninitialimplementationofthisframework.However,withthisnaiveapproachormanualdebugging,developers'wrongestimationofwhetherthefaultystatementisexecutedbeforethecheckingpoint(breakpoint)maymakethedebuggingprocessfail.Soweproposeanotherrobustapproachbasedonthisframework,handlingcaseswheredevelopersmakemistakesduringthefault-localizationprocess.Weperformedtwoexperimentalstudiesandtheresultsshowthatthetwointeractiveapproachesarequiteeffectivecomparedwithexistingfault-localizationapproaches.Moreover,therobustapproachcanhelpdevelopersfindfaultswhentheymakewrongestimationatsomecheckingpoints.
简介:Itisknownthatcriticalpathtestgenerationmethodisnotacompletealgorithmforcombinationalcircuitswithreconvergent-fanout.Inordertomadeitacompletealgorithm,weputforwardareconvergent-fanoutorientedtechnique,theprincipalcriticalpathalgorithm,propagatingthecriticalvaluebacktoprimaryinputsalongasinglepath,theprincipalcriticalpath,andallowingmultiplepathsensitizationifneeded.Relationshipamongtestpatternsisalsodiscussedtoacceleratetestgeneration.
简介:Withseveralricegenomeprojectsapproachingcompletiongeneprediction/findingbycomputeralgorithmshasbecomeanurgenttask.Twotestsetswereconstructedbymappingthenewlypublished28,469full-lengthKOMEricecDNAtotheRGPBACclonesequencesofOryzasativassp.japonica:asingle-genesetof550sequencesandamulti-genesetof62sequenceswith271genes.Thesedatasetswereusedtoevaluatefiveabinitiogenepredictionprograms:RiceHMM,GlimmerR,GeneMark,FGENSHandBGF.Thepredictionswerecomparedonnucleotide,exonandwholegenestructurelevelsusingcommonlyacceptedmeasuresandseveralnewmeasures.Thetestresultsshowaprogressinperformanceinchronologicalorder.Atthesametimecomplementarityoftheprogramshintsonthepossibilityoffurtherimprovementandonthefeasibilityofreachingbetterperformancebycombiningseveralgene-finders.