Software Engineering Economics

4IEEETRANSACTIONSONSOFTWAREENGINEERING,VOL.SE-10,NO.1,JANUARY1984SoftwareEngineeringEconomicsBARRYW.BOEHMAbstract-Thispapersummarizesthecurrentstateoftheartandyoucanthinkofanumberofdecisionsituationsinwhichyourecenttrendsinsoftwareengineeringeconomics.Itprovidesanover-hadtodeterminesomekeysoftwareproductfeatureasafunc-viewofeconomicanalysistechniquesandtheirapplicabilitytosoft-..some..wareengineeringandmanagement.Itsurveysthefieldofsoftwaretionofsomelimitingcriticalresource.costestimation,includingthemajorestimationtechniquesavailable,Throughoutthesoftwarelifecycle,1therearemanyde-thestateoftheartinalgorithmiccostmodels,andtheoutstandingcisionsituationsinvolvinglimitedresourcesinwhichsoftwareresearchissuesinsoftwarecostestimation.engineeringeconomicstechniquesprovideusefulassistance.ToIndexTerms-Computerprogrammingcosts,costmodels,manage-provideafeelforthenatureoftheseeconomicdecisionissues,mentdecisionaids,softwarecostestimation,softwareeconomics,anexampleisgivenbelowforeachofthemajorphasesinthesoftwareengineering,softwaremanagement.softwarelifecycle.*FeasibilityPhase:Howmuchshouldweinvestinin-I.INTRODUCTIONformationsystemanalyses(userquestionnairesandin-Definitionsterviews,current-systemanalysis,workloadcharacteri-zations,simulations,scenarios,prototypes)inorderThedictionarydefineseconomicsasasocialsciencethatweconvergeonanappropriatedefinitionandcon-concernedchieflywithdescriptionandanalysisoftheproduc-ceptofoperationforthesystemweplantoimple-tion,distribution,andconsumptionofgoodsandservices.ment?HereisanotherdefinitionofeconomicswhichIthinkismore.PlansandRequirementsPhase:Howrigorouslyshouldhelpfulinexplaininghoweconomicsrelatestosoftwareengi-wespecifyrequirements?Howmuchshouldweinvestneering.inrequirementsvalidationactivities(automatedcom-inresourcnimicsithesitudyiofs.hwpolmkdeionpleteness,consistency,andtraceabilitychecks,analyticinresource-limitedsituations.models,simulations,prototypes)beforeproceedingtoThisdefinitionofeconomicsfitsthemajorbranchesofdesignanddevelopasoftwaresystem?classicaleconomicsverywell.deinadevlpasfwrsytmclassicaloeconomicsverywel*ProductDesignPhase:ShouldweorganizethesoftwareMacroeconomicsiSthestudyofhowpeoplemakedecisionstaeiosbetsopeicfeitninresource-limitedsituationsonanationalorglobalscale.Itsoftareicgenerallyutncompletelyeetsnu...~~~~~~~~softwarewhichgenerallybutnotcompletelymeetsourdealswiththeeffectsofdecisionsthatnationalleadersmakeonsuchissuesastaxrates,interestrates,foreignandtraderqogremintsdpolicy.~~~''ProgrammingPhase:Givenachoicebetweenthreedatapolicy...storageandretrievalschemeswhichareprimarilyexe-Microeconomicsisthestudyofhowpeoplemakedecisionscutiontime-efficient,storage-efficient,andeasy-to-inresource-limitedsituationsonamorepersonalscale.Itdealsmodify,respectively;whichoftheseshouldwechoosewiththedecisionsthatindividualsandorganizationsmakeonsuchissusashowuchnsuancetobywhchwrdpoc-toimplement?suchissuesashowmuchinsurancetobuy,whichwordproc-*IntegrationandTestPhase:Howmuchtestingandfor-essortobuy,orwhatpricestochargefortheirproductsormalverificationshouldweperformonaproductbe-services,forereleasingittousers?EconomicsandSoftwareEngineeringManagement*MaintenancePhase:GivenanextensivelistofsuggestedIfwelookatthedisciplineofsoftwareengineering,weseeproductimprovements,whichonesshouldweimple-thatthemicroeconomicsbranchofeconomicsdealsmorewithmentfirst?*Phaseout:Givenanaging,hard-to-modifysoftwarethetypesofdecisionsweneedtomakeassoftwareengineersormanagers.~~~~~~~~~~~product,shouldwereplaceitwithanewproduct,re-ormanagers.~~~~~~~~~~~~~structureit,orleaveitalone?Clearly,wedealwithlimitedresources.ThereisneversenoughtimeormoneytocoverallthegoodfeatureswewouldOutlineofThisPaperliketoputintooursoftwareproductts.AndeveninthesedaysTollof~~~~~~~~~~~~~~chahadwran.itaeoyormrinfcn(cost-benefitanalysis,presentvalueanalysis,riskanalysis,etc.)softwareproductsmustalwaysoperatewithinaworldoflim-itedcomputerpowerandmainmemory.Ifyouhavebeeninthesoftwareengineeringfieldforanylengthoftime,IamsurelifEconomicprinciplesunderlietheoverallstructureofthesoftwareliecycle,anditsprimaryrefinementsofprototyping,incrementalde-velopment,andadvancemanship.TheprimaryeconomicdriveroftheManuscriptreceivedApril26,1983;revisedJune28,1983.life-cyclestructureisthesignificantlyincreasingcostofmakingasoft-TheauthoriswiththeSoftwareInformationSystemsDivision,warechangeorfixingasoftwareproblem,asafunctionofthephaseTRWDefenseSystemsGroup,RedondoBeach,CA90278.inwhichthechangeorfixismade.See[11,ch.4].0098-5589/84/0100-0004$O1.O0©1984IEEEBOEHM:SOFTWAREENGINEERINGECONOMICS5MASTERKEYAsindicatedinFig.1,standardoptimizationtechniquesTOSOFTWAREENGINEERINGECONOMICScanbeusedwhenwecanfindasinglequantitysuchasdollarsDECISIONANALYSISTECHNIQUES(orpounds,yen,cruzeiros,etc.)toserveasauniversalsol-ventintowhichallofourdecisionvariablescanbeconverted.AlL\USFSTANDARD|/OUTCOMEOr,ifthenondollarobjectivescanbeexpressedasconstraintsDECISIONCRITIRIAYEFSO)PTIMIZATION,ODECISIONODC)CONVERTIBLETNETVALUEHNIGHLYSEjNSITIVE(systemavailabil