a self-proclaimed voice of Project Management 2.0 claims that traditional projects management and their tools were created to support waterfall project management style. Here's a debunking of that notion
Can this really be true?
The self proclaimedleadingvoice of PM2.0 says...
Traditionalprojectmanagementsoftwareapplications,likeMS Project,werecreated to supportthe
waterfallprojectmanagementstyleand arefile-based.All thedata on differentprojectsarestored in
variousdisconnected filesand are usually accessible to the team membersin theread-only mode.The
existing combination of processesand toolsdoesnotencouragetheteamto contributeto projectplans
directly on a daily basis.With these solutions,someonehasto connectall the pieces and bitsof
information into a bigger picture,and this person is the projectmanager.Traditionalproject
managementapplicationsalso arerarely suitablefordistributed teamsthat work in a heterogeneous
environmentof multipleoperating systems.Thissoftwareisfocused on the projectmanagerand places
him or her in the center of the projectcommunications.Itoften meansthattheprojectmanagermust
collect all the dataand manually puttheinformation into the projectplan.
Thisof course mightpossiblybe anopinionbasedonnarrow experience.
But FirstDoes Waterfall = Design,Code,Test?
Firstlet'sestablishacontextforthe notionof "waterfall."Inthe redherringapproachto waterfall,a
projectisa seriesof tasks - design,code,test.Everyone I've everycome incontactwith,frommyfirst
software developmentjobasa graduate studentwritingFOCAL(aninterpretative languageforthe PDP-
15), to my lastactual codingposition,inthe late 80's writingADA forthe rudderpositionholdingunder
side pressure control loopforthe 668 classsubmarine,noone everproducedaworkingproductby
doingdesign,code,test.I'dsuggestthose thatexecute software developmentinthe design,code,test
mode are foolishatbest,andincompetentatworse.IthappensIknow - or at leastthere are storiesof it
happening.Butthatdoesn'tmeanitright.Remember Pauli andhisrightandwrongquote.
All developmentwasinincrements.Write alittle,testalittle,establishaworkingbaseline,take abreak
to thinkaboutthe nextsteps.Move to the nextstep.Whywas thisthe case inthe domainI worked?
Because inthe embeddedsystems,of radar,sonar,real time control - at leastinthose days- whenthe
CPU encounteredanerror,itstoppedrunning."The runlightwentout"inPDP-15 and 2901 bitslice
tape reader,and thenassembled,compiled,orsimply"run"toproduce the desiredresults.Youwent
slowly,stepbystep,andknewwhateverychange wouldproduce.If youdidn'tyou'dbe stuck - deadin
No more Waterfall comparisons - Let'sget to the real Pointof Managing Projects
All projectsare a sequence of workactivities.These workactivities,themselvescanbe performedin
parallel orinseries,dependingonseveral constraints:
The availabilityof resources
The logical dependenciesof the intermediate productsof the projects
In the embeddedsystemsworld,Ican't developthe control algorithminthe absence of the I/Odrivers
fr the sensorarray. I can't developthe interrupthandlingcode forthe sensorarray,in the absence of
the operatingsystemschedulingalgorithm.The functionalityof the rudderholdingcontrol loophasto
be builtina specificorder- a sequence.A Waterfallof capabilities.Notdesign,code,test.Onlyafool
woulddothat. Andhopefullyonlyonce.No,the orderof the code mustfollow asequence.
Whenasked"what'sthe purpose of time?"Einsteinsupposedlysaid
Time keeps everything from happeningatonce
So let'sstopusingwaterfall asthe anti-Christof projectmanagement.Allprojectactivitiesoccurinsome
sequentialmanner- ina series of workactivities.
Deconstructing the PM 2.0 Opinion
So let'ssetsome context.Iwas an earlyuserof MSFT Project.NotVersion1.0,butVersion3 forDOS in
1986. The official Version1for Windowswasonmy desktopmachine runningunderWindows3.1.A
complete piece of crapcomparedto TimeLine,whichranunderDOSand made beautiful picturesof the
projectswe lookedafter.Hardware developmentusingSun-1CardsrunningUnix,inthe dayswhenyou
couldcall Brian Kernighanandaskquestions,andhe'd sendpatcheson9-track tape
So the responsesbelowcome fromanecdotal evidence of havingwalkedthe walkforsome time -
Traditional projectmanagement softwareapplications,likeMSProject, were created to
supportthe waterfall project managementstyle andare file-based.
The designof MSFT projectisindependentof itsuse.Waterfall of course isthe code wordfor non-agile.
An uninformedopinionof course.The notionof Waterfall - design- code - testisforbiddeninthe US
DoD. Nodoubt,like all processes,thereare misapplications.Inthe same waythere are misapplications
of Scrumand XP.
The plan,schedule,andcostbaseline foraprojectisnot heldisa seriesof emails,tweets,IMmessages.
It isa database - the Performance MeasurementBaseline.Todootherwise,foranythingotherthana
trivial project,wouldbe likeplanningthe constructionof yourhome (anactivityI'll neverwanttodo
again),usinga bunchof stickynoteson the dashboardof the constructionsuperiorstruck.Itcan be
done,butthe resultsare usuallydisappointingtoall.
MSFT Projectis usedina wide varietyof projectmanagementprocessenvironments.Rangingfor
productionsequential effortstoScrumbasedsoftware development.
My favorite anecdote aboutusingMSFTProjectforonlywaterfall projectsstartsrighthere inBoulder.
The RallySoftware MicrosoftProjectconnector.
All the dataon different projectsare stored in variousdisconnectedfilesandare usually
accessibleto the team members inthe read-onlymode.
Thisis course issimplyBADprojectmanagement.Projectmanagementpossiblyperformedbynaive and
inexperiencedprojectmanagers.MSFTProjectisa "data base,"accessible throughavarietyof APIs.VBA
for project,VBA forAccess,andVBA for Excel.All provide integratedmanagementof the dataheld
inside MSFTProject.Grantedthere are limitsonthe fieldtypesandthe numberof fields.Thishasbeen
improvedin2007 and possibly2010.
But no credible projectmanagerwouldspreaddatainvariousdisconnectedfilesaccessible ona"read
only"basis.StartsimplybyinstallingSharePoint(WSS) andmove ontoMOSS. Buy andinstall one of
variousEnterprise ProjectManagementtoolsbasedonSharePoint.Use SafranforProject.Move up to
SAP,PeopleSoft,andOracle MSFTProjectconnectors.Possiblyeven wInsight.
We hada wonderful postercampaignata verylarge nuclearweaponsdecommissioningsite,where I
leadone of the manyProgram ManagementOffices.
Don't dostupidthingson purpose
Everytime I hearsome "expert"onprojectmanagement,speakaboutwhatisessentially"doingstupid
things,"Ithinkof those days whenstupidthingsgotpeople killed,suspended, orfired.
The existingcombinationofprocessesandtools doesnot encourage the team to contribute to
project plansdirectly ona dailybasis.Withthese solutions,someonehasto connect all the
pieces andbits of informationintoabigger picture,and thisperson is the project manager.
Thisis utternonsense inanycredible projectmanagementprocess.Whywouldanysensible project
managerprohibitcontributionstoprojectplans.Maybe noton a dailybasis.Youhave to tune the
managementof the projecttothe rhythm of the project.Thisis once againone of those "stupidthings"
comments.Basedeitheroninexperience, naivety orsimplyignorance of whatprojectmanagersactually
do fora living.
The NASA,AirForce, NavAir,andArmy programs we workmandate weeklyearnedvalue management.
EveryThursdaywe have a sitdownreview withthe Control AccountManagers(CAM) toassessprogress
for the week,plansfornextweek,nextdeliverables,nextrollingwave.Theseprocessare generally
applicable toall enterprise classprojectsnomatterthe domainorcontext.The businessrhythmisbuilt
aroundthe weeklyEV process.
At the nuclearweaponsplantdecommissioning($11Bover7 years),we hadPlanof the Day duringthe
lastyear to assure we stayedonschedule,onbudget,anddidn'tkill anyone inthe process.
The people whoconnectall the pieces,are the same people whoare willinglyaccountable fordelivering
all the pieces.The CAMs,Technical Leads,the workstreammanagers,the ProgramPlanningand
Controlsstaff,the ProgramManager and hisdeputies.There can'tbe successful anyother way.
Traditional projectmanagement applicationsalsoarerarely suitablefordistributedteams
that work in a heterogeneous environmentofmultipleoperatingsystems. This softwareis
focusedon the project manager andplaces himor her in the center ofthe project
communications.Itoftenmeans that the project manager must collectall the data and
manuallyputthe informationintothe project plan.
Nonsense again."rarelysuitable?"Gotanyphysical evidence?Mostdefense andspace programsuse
distributedteams,multipleIntegratedProjectTeams(IPT),distributedsites,andwide varietiesof tools,
environments,andworse - managementprocesses.Itsimplycan'tgetdone any otherway.
There aren't enoughpeopleinasingle buildingtoproduce the productsorservices.The toolsthat
supportthe Program Managementpieces - notthe social networking - are enterprisegrade.SAP,Oracle
for Construction,MicrosoftEnterpriseProjectManagement,specializedenterprisetoolslike wInsight.
All collaborativeweb-basedtoolstailorabletothe needsof the user.
I getthe sense thatthose wantingusto move tothe PM2.0 paradigmdon'tget out muchto see how
professionalprojectmanagersdotheirjobusingthe currenttool sets.