ohjelmistoprojektit€eivÄt voi€enÄÄ epÄonnistua! … · s2 d1 business€and€feasibility...
TRANSCRIPT
OHJELMISTOPROJEKTIT EIVÄTVOI ENÄÄ EPÄONNISTUA!
Ketterän projektinhallinan perusperiaatteitaPekka Abrahamsson
VTT TECHNICAL RESEARCH CENTRE OF FINLAND22.8.2007, Projektinhallintapäivä, Tampere
VTT TECHNICAL RESEARCH CENTRE OF FINLAND
23.8.2007 (C) VTT, Pekka Abrahamsson 2
A fact
”Universal yardstick to measuresuccess does not exist”
VTT TECHNICAL RESEARCH CENTRE OF FINLAND
23.8.2007 (C) VTT, Pekka Abrahamsson 3
A KEY PROBLEM
Who is involved? Speaker Listener Chairman Technical help Conference organizer
So, when is this presentation a success?
If we do not clarify the concept nor agreeon the terms, success remains as a subjective feeling!
VTT TECHNICAL RESEARCH CENTRE OF FINLAND
23.8.2007 (C) VTT, Pekka Abrahamsson 4
COMMON SUCCESS DIMENSIONS
Project efficiency
Benefit to the customer
Business success
Preparing the future
Source: Shenhar et al 1997
VTT TECHNICAL RESEARCH CENTRE OF FINLAND
23.8.2007 (C) VTT, Pekka Abrahamsson 5
Success as a multidimensional construct
Theoretically, what is the relationship betweenthe dimensions and the construct ?
a) Latent model
Job satisfactionJob motivation
Role stressUpwardinfluencetactics
b) Aggregate model
VTT TECHNICAL RESEARCH CENTRE OF FINLAND
23.8.2007 (C) VTT, Pekka Abrahamsson 6
Measuring upwardinfluence tactics
It is a latent model
upwardinfluencestrategies
Hard
Soft
Rational assertivenessrationality
coalitionupward appeal
ingratiationexchange
item 1item 2item 3
(Farmer etal. 1997)
VTT TECHNICAL RESEARCH CENTRE OF FINLAND
23.8.2007 (C) VTT, Pekka Abrahamsson 7
Measuring software project successIt is an aggregate model
Successin sw
project
ProjectEfficiency
BenefitFor customer
Businesssuccess
PreparationFor future
Resource useTime constraint…
One dimension does notequally manifestsuccessConstruct is formed asalgebraic functions ofits dimenions
item measure 1item measure 2
VTT TECHNICAL RESEARCH CENTRE OF FINLAND
23.8.2007 (C) VTT, Pekka Abrahamsson 8
AGILE PROJECTMANAGEMENT
VTT TECHNICAL RESEARCH CENTRE OF FINLAND
23.8.2007 (C) VTT, Pekka Abrahamsson 9
A PRINCIPLE AGILE SOLUTION
Contents
Resources Time
Contents
Resources Time
Product management steering group
VTT TECHNICAL RESEARCH CENTRE OF FINLAND
23.8.2007 (C) VTT, Pekka Abrahamsson 10
Principles of agile software development
12. Team reflectsregularly where
and how toimprove
11. The bestresults emerge
from selforganizing teams
10. Simplicity isessential
9. Continuousattention to
technicalexcellence and
good design
8. Promotesustainable
developmentpace
7. Workingsoftware is the
primary measureof progress
6. Placeemphasis onfacetoface
communication
5. Build projectsaround
motivatedindividuals
4. Businesspeople and
developers worktogether daily
throughout theproject
3. Deliverworking software
frequently
2. Welcomechanging
requirementseven late in the
project
1. Satisfycustomer
through earlyand frequent
delivery
Agilemanifesto
Practices
VTT TECHNICAL RESEARCH CENTRE OF FINLAND
23.8.2007 (C) VTT, Pekka Abrahamsson 11
COMMON BELIEFS OFAGILE DEVELOPMENT
•Agile software development leads to hacking•Agile means no documentation•We have problems, agile is the easy way out•Agile is best suited for small team development•There is very little room for architecture work in
agile development•Agile means that we do not think about the
future•Agile means that we do not need to plan•Agile means incremental & iterative; we have
always been incremental and iterative > thereis nothing new here
VTT TECHNICAL RESEARCH CENTRE OF FINLAND
23.8.2007 (C) VTT, Pekka Abrahamsson 12
Change is the only certainty insoftware design & development
The production of Technical Specifications for a 3rd GenerationMobile System based on the evolved GSM core networks.
VTT TECHNICAL RESEARCH CENTRE OF FINLAND
23.8.2007 (C) VTT, Pekka Abrahamsson 13
What is Agility?
•Agility, for a software development organization,is the ability to adapt and react expeditiously*and appropriately to changes in its environmentand to demands imposed by this environment”[Kruchten, 2001]
* expeditiously, d. nopeasti, vauhdikkaasti,joutuisasti
•Thus, agility is the ability to both create andrespond to change in order to profit in a turbulentbusiness environment
•Agile organizations are•nimble, i.e. able to change directions quickly•flexible, i.e. able to change the way things
are done
VTT TECHNICAL RESEARCH CENTRE OF FINLAND
23.8.2007 (C) VTT, Pekka Abrahamsson 14
What Kinds of ProblemsDoes Agility Solve Best?
•Complex problems characterized by•change,•speed, and•turbulence.
•Situations where the development is required•to meet the tight delivery schedules, and•to reduce the significant risk and uncertainty
that generate constant change during theproject.
•Change is imposed by rapidly evolvingtechnology, business and product needs
VTT TECHNICAL RESEARCH CENTRE OF FINLAND
23.8.2007 (C) VTT, Pekka Abrahamsson 15
AGILE SOFTWARE DEFINED
•Technical Excellence•Cooperation•Timeboxed development•Fixed costs & delivery date•Continuous risk management
VTT TECHNICAL RESEARCH CENTRE OF FINLAND
23.8.2007 (C) VTT, Pekka Abrahamsson 16
IN ORDER ANYTHING TO HAPPEN THEDEVELOPERS NEED TO BE ONBOARD!
Source: Nokia, publicly available at:http://www.odde.com/articles/2006/nokia_agile.pdf
VTT TECHNICAL RESEARCH CENTRE OF FINLAND
23.8.2007 (C) VTT, Pekka Abrahamsson 17
"PROVEN" BENEFITS OF AGILE DEV.:THE PROJECT, PRODUCT MANAGERS & BUSINESS
•Improved control over product and/or service> You're truly in the driver's seat> Every 24 weeks you decide> Continuous risk management> Accept losses or declare victory at any
given control point•Visibility of projects improve by several
magnitudes > lesser need for status reporting•Improved focus on real issues: You concentrate
on product roadmap, future and communication
VTT TECHNICAL RESEARCH CENTRE OF FINLAND
23.8.2007 (C) VTT, Pekka Abrahamsson 18
SCRUM FOR AGILEPROJECT
MANAGEMENT
VTT TECHNICAL RESEARCH CENTRE OF FINLAND
23.8.2007 (C) VTT, Pekka Abrahamsson 19
Scrum Roles & Responsibilities
Support Other
Management Customer / Business Competence
Develo
pmentTeam
Scrum MasterEnsures the project runsunder Scrum principles,runs the Scrum Team
ManagementIn charge of finaldecision making
Scrum TeamProject team, with
authority to decide onactions
CustomerDefines the set of requirements
for the system (the ProductBacklog list)
Product OwnerResponsible for the project, and therealization of the requirements into
features
VTT TECHNICAL RESEARCH CENTRE OF FINLAND
23.8.2007 (C) VTT, Pekka Abrahamsson 20
FSecure's starting point (2004)
DiscontinuationGeneral AvailabilityProduct Realization
R1S1 V3 V1V2
ReleaseDevelopment
D1S2
Business and FeasibilityStudy
SystemTest
BetaValidation
RCValidation ReleasingDevelopment Iterations
Screening Validation
D2
Product lifecycle and product realization cycle
DA Dn ...
Product &Project
Elaboration
Product &Project
Initiation
Product Life Cycle Management
Source: AgileITEA newsletter 1/2006
VTT TECHNICAL RESEARCH CENTRE OF FINLAND
23.8.2007 (C) VTT, Pekka Abrahamsson 21
FSecure's new production model FLEX (2006)
Source: AgileITEA newsletter 1/2006
VTT TECHNICAL RESEARCH CENTRE OF FINLAND
23.8.2007 (C) VTT, Pekka Abrahamsson 22
•Due to the adoption of an agileproduction frame, FSecure's mobiledivision in a new product developmentachieved§ 3x reduction in leadtime,§ 4x cheaper,§Best quality ever
•Innovative leap due to the acceptance ofradical variation in product development& management processes
THE CASE OF AN INNOVATIVE LEAP
Fact corner: SME of 300 developers Mobile & desktop sw Products sold globally
"From the management point of view, the results wereamazing." Jari Still, Director, FSecure, Finland
Source: AGILEITEA Newsletter #2, 2005AGILEITEA Newsletter #1, 2006
(http://www.agileitea.org)
VTT TECHNICAL RESEARCH CENTRE OF FINLAND
23.8.2007 (C) VTT, Pekka Abrahamsson 23
QUANTIFIED QUALITY "CONTROL"
8,6 %
9,7 %35,1 %
51,6 %
26,7 %
22,9 %16,4 %
7,8 %
0,0 %
1,8 %
0,0 %8,6 %
0 %
10 %
20 %
30 %
40 %
50 %
60 %
1 2 3 4 5
Release
Per
cem
tage
of t
otal
effo
rt
Defect fixEnhancementNew feature
RELEASE
VTT TECHNICAL RESEARCH CENTRE OF FINLAND
23.8.2007 (C) VTT, Pekka Abrahamsson 24
20,0
15,0
10,0
5,0
0,0
5,0
10,0
15,0
20,0
25,0
zOmbie
eXpert
bAmbie
Bubble
uniCorn
ESTIMATION PRECISION IN AGILE DEVELOPMENT(I.E. HOURS LOST BY FAULTY ESTIMATES)
Bottom line:eXpert: 92 % tasks within +/ 5 hours error marginzOmbie: 85%, bAmbie: 93%, uniCorn: 89%, Bubble: 69%
VTT TECHNICAL RESEARCH CENTRE OF FINLAND
23.8.2007 (C) VTT, Pekka Abrahamsson 25
Agile productivityCase Philips: Modena project
•Industry data:•Average: 150 SLOC/MM•Bestinclass: 750 SLOC/MM•PSP: 4000 SLOC/MM
•Case agile (Modena project)•New code only: 858 SLOC/MM•New + (0.3 * Adapted): 1143 SLOC/MM•New + Adapted: 1809 SLOC/MM
Based on total effort including Management &external Test effort
Source: AGILEITEA Newsletter #1, 2006
VTT TECHNICAL RESEARCH CENTRE OF FINLAND
23.8.2007 (C) VTT, Pekka Abrahamsson 26
INNOVATIVE LEAP EXPLAINED
•"The actual processes adopted were neither novelnor particularly inventive. Rather they had thevirtues of being easy to explain and relatively easyto comply with, with goals easily describable ashaving been met or not."
Ward et al. (2001)
VTT TECHNICAL RESEARCH CENTRE OF FINLAND
23.8.2007 (C) VTT, Pekka Abrahamsson 27
THANK YOU!
Questions and comments?
Contact me at:[email protected]. 0405415929