Using Quality Criteria in Programming Industry: A Case Study

UsingQualityCriteriainProgrammingIndustry:ACaseStudyMartin Blom, Eivind J.Nordby,Donald F.RossUniversityof Karlstade-mail:Martin.Blom@hks.seUniversityof Karlstad651 88 Karlstadfax:+46 54 83 84 46Erland Jonssone-mail:jonsson@ce.chalmers.seChalmers Universityof TechnologyInstitutionen for datorteknik,Horsalsvagen11412 96 Goteb orgJune16,1998Neither this pap er nor anyversion closeto it hasb eenor is b eing o eredelsewherefor publication.All necessary clearances have b een obtained for the publication of this pap er.If accepted,the pap er willb e made available in Camera-ready forms by June 15th,1998,and it will b e p ersonally presented at theEUROMICRO 98 Conference by the author or one of the co-authors.The presenting author(s) will pre-register for EUROMICRO 98 b efore the due date of the Camera-ready pap er.Martin BlomAbstractSemanticasp ectsofprogramminghaveb eenstudiedinacademiaformanyearsasameanstoachievehighersoftwarequality.Itisnotclearwhichextentthesehaveb eenacceptedinprogrammingindustry.Therefore,wmadeacasestudyofan industrial pro ject using a checklist of quality criteria sp eci cally pro duced for thisstudy.Theresultsfromthe casestudy showthathighleveldo cumentation and lolevel do cumentation is sucient.However, there no semantic description constituting agap b etween these to.This gap, which constitutes a p otential source of errors, mustb e lled in order to signi cantly improve software quality.1Intro ductionIn the early years of programming the main concern was to develop syntax, i.e.languageswith whichyou could communicate with the computer, but over the years the concern hasmovedoertosemanticissues[2,6].Inacademia,theoriesforcorrectnesshasb een an imp ortant issue for manyyears [3, 7].We b elieve, however, that there is a lackofunderstanding of semantic asp ects in programming industry.To see if this assumption holds,a case study of an industrial pro ject was done.1