CMMI_Agile_Webinar

©2008CarnegieMellonUniversityCMMI&AgileSoftwareEngineeringInstituteCarnegieMellonUniversityPittsburgh,PA15213MikeKonradShaneMcGrawSEIWebinarNovember13,20082©2008CarnegieMellonUniversityBriefBiosofSEIStaffMikeKonrad•WithSEIfor20years•CMMIArchitect,CCBChair,ModelTeamLead•Yourpresenter!ShaneMcGraw•WithSEIfor6years•SPINCoordinatorfortheSEI•Yourhost!3©2008CarnegieMellonUniversityTheProcessImprovementLandscapeCMMIandAgileArePerceivedtobeatOddswithEachOtherReliantonprocessdefinitions,measures,andartifactsPredictiveGenerallydrivenbymanagementSkepticalofprocessdefinitions,measures,andartifactsEmergentOftenbegunasgrass-rootseffortCMMIandAgileproponentsareapttodemonizeorignoreeachother.4©2008CarnegieMellonUniversitySEIReportonCMMI&AgileTobuildawarenessofthesynergies,theSEIhasco-authoredareport:CMMIorAgile:WhyNotEmbraceBoth!’sco-authorsincludeexpertAgilepractitionersintheAgilecommunity:•HillelGlazer–authorofoneofthefirstarticlesonCMMandAgile(2001)andHillelGlazer’sBlogat•JeffDalton–co-authoroftheforthcomingAgileCMMI(tobepublishedbyAuerbach)andauthorofJeffDalton’sBlog:AskTheCMMIAppraiserat•DaveAnderson–co-authorofTheDeclarationofInterdependence,programmanagerforMSFforCMMI,authorofoneofthefirstbooksonanAgilemethodology(FDD),andauthorofDaveAnderson’sWebsiteandBlogat©2008CarnegieMellonUniversityReasonsforDiscordEarlyadoptersofeachapproachrepresentextremes:•CMMI:Large-scale,missioncritical,extensiveoversight•Agile:Fast-moving,empoweredteams,volatilerequirements,accesstocustomerNegativeperceptionsoftheotherside•20yearsofCMMandthenCMMIsometimesbeingmisapplied•Top-downvs.bottom-upapproaches•Terminologyandstylecanbe“turnoffs!”–Technicaldatapackage,predictability,ExtremeprogrammingLackofAccurateInformation•AgilistasmorefamiliarwithCMM•ManyincrementalimprovementsinCMMIgounrecognized•SomeclaimingtobeAgile(orusingCMMI)arenot!•AgileDOESrequiresdiscipline6©2008CarnegieMellonUniversityCMMIUsersNeedtoKnowAgilearose,inpart,asabacklashtoCMMandCMMIMisuseAgileappliessoundsSWEprinciplesincertainprojectcontextsPractitionersneedtobeinvolvedinimplementingCMMI•Addressbusinessneedsandpriorities•Capturewhatisalreadydonewell•Useunderstandablelanguageandformalisms•Balancestandardizationandtailoring•Ensureandsupportdeploymenttonewprojects(OPFSG3,IPMSP1.1)•Maketheprojectprocesstransparent•Encourage,evaluate,andimplementfeedbackandimprovementsuggestions•Periodicallyrevise•Articulatebenefits7©2008CarnegieMellonUniversityCMMIUseShouldBeMoreLiketheLeftAppropriateuseMisuseFocusonimprovedperformance;ratingsshouldbeanaturalby-productFocusonmaturitylevelUsetoidentify,motivate,learnabout,andimproveprocessesbywhichthebusinessgetsdoneUseasaprocessstandard(Note:CMMIsaysit“containsneitherprocessesnorprocedures”)Usetoachievegreaterflexibilityandleannessindynamic,high-trustenvironmentsUsetocreatebureaucracy;unnecessarilyceremonial,wastefulprocessesRegularlymigratetonewerversionsofCMMI;investigatingwhatchangedandwhySticktotheolderversions;failtoinvestigatereasonsforchange;fallingbehind(andgivingrealCMMIpractitionersabadname)Appropriatelycomplementingthejourneywithnewframeworks,methods,andtechnologies;learnwhatworksanduseit!OnlystickingtoCMMIandignoringordemonizingSixSigma,Agile,etc.ConsidertheinformativematerialininterpretingandimplementingthepracticesEitherignoringtheinformativematerialorusingitasachecklist8©2008CarnegieMellonUniversitySkillinCMMIInterpretationisNecessaryGoalsarerequired,notpracticesPracticesareexpected,notsubpractices,typicalworkproductsButinformativematerialisstillimportantasitclarifiespracticeintent•apracticestatementisasummaryofthelargerconceptcharacterizedbytheinformativematerialaroundit•supportscorrectinterpretationandimplementation•helpsunderstanddependencieswithotherpartsofmodel•describesimplementationtypicaloflarge,complexprojects,highcostoffailureFailingtounderstandtheaboveeitherleadstosuperficialprocessesorover-prescriptivestandardprocesses(causingwasteandfrustration)Instead,eachlevel(required,expected,informative)servesasthestartingpointforthelevelabove•HowimportantisthatPAgoaltoachievingthebusiness’sobjectives?•Whatactivitiesarenecessarytoachievingthegoal?9©2008CarnegieMellonUniversityWhatAgileAddressesthatCMMIDoesn’tLightweightprovenapproachesthatworkwellforsmallco-locatedteamswithembeddedcustomer(e.g.,XP)•Insomecases,largerteamsaswell(FDD,more-recentScrumandTSPvariants,andmorerobustversionsofCrystal)Asetof“howto’s”tostandardizeandbuildaprocessinfrastructurearound(OSSP,PAL,etc.)ProvidingpractitionersmorevoiceintheprocessestheyuseProvidingcustomersmorevoiceintheproductsandservicestheyreceive10©2008CarnegieMellonUniversityImprovementsMedian#ofdatapointsLowHighCost20%213%87%Schedule37%192%90%Productivity67%1611%255%Quality50%1829%132%CustomerSatisfaction14%6-4%55%ReturnonInvestment4.8:1142:127.7:1•N=25,asof15December2005•Organ