SERA 2003Keynote AddressResearch Review

San Francisco, CAJune 25, 2003

A forthcoming Addison Wesley Book

©USC-CSE6/25/03 1

6/25/03 ©USC-CSE 2

Background

• Two approaches to software development – Disciplined (SW-CMM, document-based, heavy

process)– Agile (XP, tacit knowledge, light process)

• Both have strengths and weaknesses• Agile and disciplined proponents are

believers• Many of us are perplexed

6/25/03 ©USC-CSE 3

The Agile Manifesto - I

• Individuals and interactions over processes and tools• Working software over comprehensive documentation• Customer collaboration over contract negotiation• Responding to change over following a plan

We are uncovering better ways of developing software by doing it and helping others do it.Through this work we have come to value:

That is, while there is value in the items on the right, we value the items on the left more.

6/25/03 ©USC-CSE 4

The Agile Manifesto – II

• Our highest priority is to satisfy the customerthrough early and continuous deliveryof valuable software.

• Welcome changing requirements, even late in development. Agile processes harness change for the customer's competitive advantage.

• Deliver working software frequently, from a couple of weeks to a couple of months, with a preference to the shorter timescale.

• Business people and developers must work together daily throughout the project.

6/25/03 ©USC-CSE 5

The Agile Manifesto – III

• Build projects around motivated individuals. Give them the environment and support they need, and trust them to get the job done.

• The most efficient and effective method of conveying information to and within a development team is face-to-face conversation.

• Working software is the primary measure of progress.

• Agile processes promote sustainable development. The sponsors, developers, and users should be able to maintain a constant pace indefinitely.

6/25/03 ©USC-CSE 6

The Agile Manifesto – IV

• Continuous attention to technical excellence and good design enhances agility.

• Simplicity – the art of maximizing the amount of work not done – is essential.

• The best architectures, requirements, and designs emerge from self-organizing teams.

• At regular intervals, the team reflects on how to become more effective, then tunes and adjusts its behavior accordingly.

6/25/03 ©USC-CSE 7

Various Agile Methods Available

• Adaptive Software Development (ASD)• Agile Modeling• Crystal methods• Dynamic System Development Methodology

(DSDM)* eXtreme Programming (XP)• Feature Driven Development• Lean Development• Scrum

6/25/03 ©USC-CSE 8

XP Principles – I

• Philosophy: Take known good practices and push them to extremes

• “If code reviews are good, we’ll review code all the time”

• “If testing is good, we’ll test all the time”

• “If design is good, we’ll make it part of everybody’s daily business”

6/25/03 ©USC-CSE 9

XP Principles – II

• “If simplicity is good, we’ll always leave the system with the simplest design that supports its current functionality”

• “If architecture is important, everybody will work defining and refining the architecture all the time”

• “If integration testing is important, then we’ll integrate and test several times a day”

6/25/03 ©USC-CSE 10

XP Principles – III

• “If short iterations are good, we’ll make the iterations really, really short – seconds and minutes and hours, not weeks and months and years”

• “If customer involvement is good, we’ll make them full-time participants”

6/25/03 ©USC-CSE 11

XP: The 12 Practices

• The Planning Game• Small Releases• Metaphor• Simple Design• Testing• Refactoring

• Pair Programming• Collective Ownership• Continuous Integration• 40-hour Week• On-site Customer• Coding Standards

-Used generatively, not imperatively

6/25/03 ©USC-CSE 12

Counterpoint: A Skeptical View – ILetter to Computer, Steven Rakitin, Dec.

2000

“individuals and interactions over processes and tools” Translation: Talking to people gives us the flexibility to do whatever we

want in whatever way we want to do it. Of course, it’s understood that we know what you want - even if you don't.

“working software over comprehensive documentation” Translation: We want to spend all our time coding. Real programmers

don’t write documentation.

6/25/03 ©USC-CSE 13

Counterpoint: A Skeptical View – IILetter to Computer, Steven Rakitin, Dec.

2000

“customer collaboration over contract negotiation” Translation: Let's not spend time haggling over the details, it only

interferes with our ability to spend all our time coding. We’ll work out the kinks once we deliver something...

“responding to change over following a plan” Translation: Following a plan implies we would have to spend time

thinking about the problem and how we might actually solve it. Why would we want to do that when we could be coding?

6/25/03 ©USC-CSE 14

Sources of Perplexity

• Distinguishing method use from method misuse– Claiming XP use when simply “not documenting”– “CMM Level 4 Memorial Library” of 99 2-inch binders

• Overgeneralization based on the most visible instances– XP is Agile; CMM is Discipline

• Multiple definitions– Quality: customer satisfaction or compliance?

• Claims of universality– The pace of IT change is accelerating and Agile

methods adapt to change better than disciplined methods therefore Agile methods will take over the IT world.

– Software development is uncertain and the SW-CMM improves predictability therefore all software developers should use the SW-CMM

6/25/03 ©USC-CSE 15

Sources of Perplexity (2)

• Early success stories– Chrysler project that successfully birthed XP was

later cancelled– Cleanroom has never made it into the mainstream

• Purist interpretations (and internal disagreements)– “Don't start by incrementally adopting parts of

XP. Its pieces fit together like a fine Swiss watch“– "An advantage of agile methods is that you can

apply them selectively and generatively."– "If you aren't 100% compliant with SW CMM Level

3, don't bother to bid" – "With the CMMI continuous interpretation, you can

improve your processes in any order you feel is best."

6/25/03 ©USC-CSE 16

Key Definitions

• Agile method – – one which fully adopts the four value propositions

and twelve principles in the Agile Manifesto.• Discipline – (per Webster)

– control gained by enforcing obedience or order; – orderly or prescribed conduct or pattern of

behavior; – self-control.

• Plan-driven – – a description for disciplined methods (order is

often defined in plans) • Plan – (per Webster)

– a method for achieving an end (a process plan); – an orderly arrangements of parts of an overall

design (a product plan). – We assume that such plans are documented.

6/25/03 ©USC-CSE 17

Finding Middle Ground

• A pragmatic view• Both approaches have “home grounds”• A broad range of environments and needs• Trends require better handling of rapid

change• Processes should be the right weight for the

job• We believe risk-based analysis is the key to

finding middle ground

6/25/03 ©USC-CSE 18

Management Characteristics

• Customer Relations– A: Dedicated, collocated– D: Contractual– A: Trust through working software and

participation– D: Trust through process maturity evaluations

• Planning and Control– A: Means to an end– D: Anchor processes, communication

• Project Communications– A: Tacit, interpersonal knowledge– D: Explicit, documented knowledge

6/25/03 ©USC-CSE 19

Technical Characteristics

• Requirements– A: Adjustable, informal stories– D: Formally baselined, complete, consistent

specifications

• Development– A: Simple Design– D: Architecture-based design

• Testing– A: Automated, test-driven– D: Planned, requirements-driven

6/25/03 ©USC-CSE 20

Technical Characteristics (2)Cost of Change

Cos

t of

Cha

nge

Time

Beck

LiLi

6/25/03 ©USC-CSE 21

Technical Characteristics (3)How Much Architecting?

Beck

LiuLiu

0

10

20

30

40

50

60

70

80

90

100

0 10 20 30 40 50 60

Percent of Time Added for Architecture and Risk Resolution

Per

cen

t of T

ime

Ad

ded

to O

vera

ll S

ched

ule

Percent of Project Schedule Devoted to Initial Architecture and Risk Resolution

Added Schedule Devoted to Rework(COCOMO II RESL factor)

Total % Added Schedule

Sweet Spot Drivers:

Rapid Change: leftward

High Assurance: rightward

Sweet spot

10 KSLOC

100 KSLOC

10,000 KSLOC

6/25/03 ©USC-CSE 22

Personnel Characteristics

• Customers– A: CRACK customers throughout development– D: CRACK customers early

• CRACK: Collaborative, Representative, Authorized, Committed, and Knowledgeable

• Developers– A: Heavy mix of high caliber throughout– D: Heavy mix early with lower capability later

• Culture– A: Many degrees of freedom (Thrives on

chaos)– D: Clear policies and procedures (Thrives on

order)

6/25/03 ©USC-CSE 23

Personnel Characteristics (2)Modified Cockburn Levels

(Ski

ll, U

nders

tandin

g)

(Formality, Documentation)

(Ski

ll, U

nders

tandin

g)

(Formality, Documentation)

Level Characteristics

3 Able to revise a method (break its rules) to fit an unprecedented new situation

2 Able to tailor a method to fit a precedented new situation

1A With training, able to perform discretionary method steps (e.g., sizing stories to fit increments, composing patterns, compound refactoring, complex COTS integration). With experience can become Level 2.

1B With training, able to perform procedural method steps (e.g. coding a simple method, simple refactoring, following coding standards and CM procedures, running tests). With experience can master some Level 1A skills.

-1 May have technical skills, but unable or unwilling to collaborate or follow shared methods.

6/25/03 ©USC-CSE 24

Summary of Home GroundsCharacteristics Agile Disciplined

Application

Primary Goals Rapid value; responding to change Predictability, stability, high assurance

Size Smaller teams and projects Larger teams and projects

Environment Turbulent; high change; project-focused Stable; low-change; project/organization focused

Management

Customer Relations

Dedicated on-site customers; focused on prioritized increments

As-needed customer interactions; focused on contract provisions

Planning/Control

Internalized plans; qualitative control Documented plans, quantitative control

Communications

Tacit interpersonal knowledge Explicit documented knowledge

Technical

Requirements Prioritized informal stories and test cases; undergoing unforseeable change

Formalized project, capability, interface, quality, forseeable evolution requirements

Development Simple design; short increment; refactoring assumed inexpensive

Extensive design; longer increments; refactoring assumed expensive

Test Executable test cases define requirements, testing

Documented test plans and procedures

Personnel

Customers Dedicated, collocated CRACK* performers CRACK* performers, not always collocated

Developers At least 30% full-time Cockburn level 2 and 3 experts; no Level 1B or -1 personnel**

50% Cockburn Level 3s early; 10% throughout; 30% Level 1B’s workable; no Level -1s**

Culture Comfort and empowerment via many degrees of freedom (thriving on chaos)

Comfort and empowerment via framework of policies and procedures (thriving on order)

* Collaborative, Representative, Authorized, Committed, Knowledgeable** These numbers will particularly vary with the complexity of the application

6/25/03 ©USC-CSE 25

Five Critical Decision Factors

• Represent five dimensions• Size, Criticality, Dynamism, Personnel,

Culture Personnel

Dynamism (% Requirements-change/month)

Culture (% thriving on chaos vs. order)

Size (# of personnel)

Criticality (Loss due to impact of defects)

5030

105

1

90

70

50

30

10

3

10

30

100

300

35

30

25

20

15

Essential Funds Discretionary

Funds Comfort

Single Life

Many Lives

(% Level 1B) (% Level 2&3)

0

10

20

30

40

Agile

Disciplined

6/25/03 ©USC-CSE 26

Two Case Studies

• Show how that agile and disciplined methods can be extended

• Provide examples of success and lessons learned

• ThoughtWorks Lease Management– Agile extended with disciplined

• CCPDS-R– Disciplined overlaid with agile concepts

6/25/03 ©USC-CSE 27

Thoughtworks Lease Management

• XP replaced ineffective traditional development• Problems when project moved beyond XP

assumptions– The effort to develop or modify a story really does not

increase with time and story number– Trusting people to get everything done on time is

compatible with fixed schedules and diseconomies of scale

– Simple design and YAGNI scale up easily to large projects

• Disciplined practices enabled XP to scale up– High-level architectural plans to provide essential

big-picture information– More careful definition of milestone completion

criteria to avoid “finishing” but not finishing– Use of design patterns and architectural solutions

rather than simple design to handle foreseeable change

6/25/03 ©USC-CSE 28

CCPDS-R

• USAF Command Center Processing and Display System Replacement for early missile warning

• Government contract environment required disciplined approach, project needed agility

• Practices implemented to provide agility mapped well to the Agile Manifesto – Individuals and Interactions over Processes and

Tools• Milestone content was redefined • Architecture was organized around the performers’ skill levels

– Working Software over Comprehensive Documentation

• Later PDR demonstrated working software for high-risk areas – Customer Collaboration Over Contract

Negotiation • Used COCOMO for cost and schedule negotiations

– Responding to Change Over Following a Plan • Automated plans• Architecture for re-use saved effort

6/25/03 ©USC-CSE 29

CCPDS-R (2) Cost of Change

RATIONALS o f t w a r e C o r p o r a t I o n

Project Development Schedule15 20 25 30 35 40

40

30

20

10

Design Changes

Implementation Changes

MaintenanceChanges and ECP’s

HoursChange

RATIONALS o f t w a r e C o r p o r a t I o n

Project Development Schedule15 20 25 30 35 40

40

30

20

10

Design Changes

Implementation Changes

MaintenanceChanges and ECP’s

HoursChangeHours

Change

Cos

t of

Cha

nge

Time

Beck

6/25/03 ©USC-CSE 30

Using Risk to Balance Discipline and Agility - Overview

Step 5. Execute and Monitor

Step 4. Tailor Life Cycle

Step 3. Architecture Analysis

Step 1. Risk Analysis

Step 2. Risk Comparison

Rate the project’s environmental, agility-oriented and discipline-

oriented risks.

Uncertain about

ratings?

Buy information via prototyping, data

collection and analysis

Compare the agile

and disciplined

risks

Go Risk-driven Agile

Agility risks dominate

Discipline risks dominate

Architect application to encapsulate agile parts

Go Risk-driven Agile in agile

parts; Go Risk-driven Disciplined

elsewhere

Yes

No

Go Risk-driven Disciplined

Tailor life cycle process around anchor point

commitment milestones

Monitor progress and risks/opportunities,

readjust balance and process as appropriate

Neither dominate

Deliver incremental capabilities according to

strategy

6/25/03 ©USC-CSE 31

Risks Used in Risk Comparison

• Environmental risks– Technology uncertainties– Many stakeholders– Complex system-of-systems

• Agility-oriented risks– Scalability– Use of simple design– Personnel turnover– Too-frequent releases– Not enough agile-capable people

• Discipline-oriented risks– Rapid change– Need for rapid results– Emergent requirements– Not enough discipline-capable people

6/25/03 ©USC-CSE 32

Three Examples

• Agent-based systems• Small – Event Managers application• Medium – SupplyChain.com application• Large – National Information System for

Crisis Management (NISCM) application• Each example results in a development

strategy based on risk analyses

6/25/03 ©USC-CSE 33

Event Managers Project

• Small startup company • Diverse set of smaller agent-based

planning applications • This project: support for managing the

infrastructure and operations of conferences and conventions

• Widening variety of options and interdependencies, makes an agent-based approach highly attractive

6/25/03 ©USC-CSE 34

Event Managers Profile

Personnel

Dynamism (% Requirements-change/month)

Culture (% thriving on chaos vs. order)

Size (# of personnel)

Criticality (Loss due to impact of defects)

5030

105

1

90

70

50

30

10

3

10

30

100

300

35

30

25

20

15

Essential Funds Discretionary

Funds Comfort

Single Life

Many Lives

(% Level 1B) (% Level 2&3)

0

10

20

30

40

Risk Items Risk Ratings Event Managers

Environmental Risks E1. Technology uncertainties E2. Many stakeholders E3. Complex system of systems Risks of using Agile Methods A1. Scalability A2. Use of simple design A3. Personnel turnover A4. Too-frequent releases A5. Not enough agile-capable people Risks of using disciplined methods D1. Rapid change D2. Need for rapid results D3. Emergent requirements D4. Not enough discipline-capable people

Risk rating scale: - minimal risk - moderate risk

- Serious but manageable risk - Very serious but manageable risk - Show-stopper risk

6/25/03 ©USC-CSE 35

Event Managers Strategy

1. StartupTeambuilding and Shared

Vision

2. Design, Development and Deployment

• Establish CRACK joint management team

• Develop shared vision, results chain, life cycle strategy, infrastructure choices

• Establish commitment to proceed, incremental capabilities, backlog

• Develop next-increment features

• Test, exercise, deploy new increment• Re-prioritize next-increment features• Analyze lessons learned; prepare strategy for next increment

Customer Management; Representative Users

Joint Developer-Customer Agile Team

6/25/03 ©USC-CSE 36

SupplyChain.com Profile

• Turnkey agent-based supply chain management systems

• Distributed, multi-organization teams of about 50 people

• Parts of applications are relatively stable, while others are highly volatile

• Architectures are driven by a few key COTS packages that are also continually evolving

6/25/03 ©USC-CSE 37

SupplyChain.com Profile

Personnel

Dynamism (% Requirements-change/month)

Culture (% thriving on chaos vs. order)

Size (# of personnel)

Criticality (Loss due to impact of defects)

5030

105

1

90

70

50

30

10

3

10

30

100

300

35

30

25

20

15

Essential Funds Discretionary

Funds Comfort

Single Life

Many Lives

(% Level 1B) (% Level 2&3)

0

10

20

30

40

Risk Items Risk Ratings SupplyChain.com

Environmental Risks E1. Technology uncertainties E2. Many stakeholders E3. Complex system of systems Risks of using Agile Methods A1. Scalability A2. Use of simple design A3. Personnel turnover A4. Too-frequent releases A5. Not enough agile-capable people Risks of using disciplined methods D1. Rapid change D2. Need for rapid results D3. Emergent requirements D4. Not enough discipline-capable people

Risk rating scale: - minimal risk - moderate risk

- Serious but manageable risk - Very serious but manageable risk - Show-stopper risk

6/25/03 ©USC-CSE 38

SupplyChain.com Strategy

Startup 1. Teambuilding and Shared

Vision

3. Design, Development and Deployment2. Systems Definition and Architecting

Stakeholders

Furnish CRACK representatives and alternates

Staff and organize to cover all success-critical risk areas

• Develop benefits-realization results chains

• Identify missing stakeholders

• Enhance mutual understanding

• Explore goals, options, technology, prototypes, risks

Review; Commit

to proceed

• Elaborate supply chain operational concept, prototypes, transition strategy

• Evaluate and determine COTS, reuse, and architecture choices

• Prioritize and sequence desired capabilities, outstanding risks

• Establish project organization, overall process, and feature teams

Review; Commit

to proceed

• Ensure representative exercise of incremental capabilities

• Monitor progress and new operational development

• Monitor project progress, risk resolution, and new technology developments

• Identify and work critical project-level actions

• Develop and integrate new feature increments

• Communicate proposed redirections

• Prepare for and execute acceptance tests, installation, operational evolution

• Analyze feedback on current version

• Reprioritize features

• Prepare strategy for next increment

• Consolidate process lessons learned

• Use risk to determine content of artifacts

Project Manager andRisk Management Team

Agile Feature Team

6/25/03 ©USC-CSE 39

NISCM Profile

• Broad coalition of government agencies and critical private-sector organizations

• Support cross-agency and public-private sector coordination of crisis management activities

• Adaptive mobile network, virtual collaboration, information assurance, and information integration technology

• Private-sector system-of-systems integration contractor

6/25/03 ©USC-CSE 40

Risk Items Risk Ratings NISCM

Environmental Risks E1. Technology uncertainties E2. Many stakeholders E3. Complex system of systems Risks of using Agile Methods A1. Scalability — A2. Use of simple design — A3. Personnel turnover A4. Too-frequent releases A5. Not enough agile-capable people — Risks of using disciplined methods D1. Rapid change D2. Need for rapid results D3. Emergent requirements D4. Not enough discipline-capable people

Risk rating scale: - minimal risk - moderate risk

- Serious but manageable risk - Very serious but manageable risk - Show-stopper risk

NISCM Profile

Personnel

Dynamism (% Requirements-change/month)

Culture (% thriving on chaos vs. order)

Size (# of personnel)

Criticality (Loss due to impact of defects)

5030

105

1

90

70

50

30

10

3

10

30

100

300

35

30

25

20

15

Essential Funds Discretionary

Funds Comfort

Single Life

Many Lives

(% Level 1B) (% Level 2&3)

0

10

20

30

40

6/25/03 ©USC-CSE 41

Startup Teambuilding and Shared

Vision

Design, Development and Deployment

Systems Definition and Architecting

Stable, Higher-criticality Module Teams

Stakeholders

Furnish CRACK representatives and alternates

• Staff and organize to cover all success-critical risk areas

• Prepare for and select competitors for concept definition

• Develop results chains• Identify missing

stakeholders• Enhance understanding• Explore goals, options,

technology, prototypes, risks

• Negotiate mutually satisfactory objectives and priorities

• Formulate top-level operational concept, requirements, architecture, life cycle plan, feasibility rational

• Select winning system integration contract

Hold Life Cycle

Objective Review. If feasibility

rationale is valid,

commit to proceed

• Prepare for/select competitors for component subcontract definition

• Elaborate operational concept, prototypes, transition strategy

• Evaluate/determine COTS, reuse, and architecture choices

• Develop/ validate critical-infrastructure software

• Prioritize/sequence desired capabilities, outstanding risks

• Formulate definitive operational concept, requirements, architecture, life cycle plan, feasibility rationale

• Use to solicit/select component subcontractors

Hold Life Cycle

Architecture Review. If feasibility

rationale is valid,

commit to proceed

• Ensure representative exercise of incremental capabilities

• Monitor progress and new operational development

• Monitor project progress, risk resolution, and new technology developments

• Identify/work critical project-level actions

• Continuously integrate/ test growing software infrastructure and components

Dynamic, Lower-criticality Module Teams

• Develop compatible component-level prototypes, requirements, architectures, plans, critical software, feasibility rationale

• Classify modules by likely stability and criticality

Organize into disciplined and agile module teams

Plan, specify, develop stable, higher-criticality module increments

Plan, specify, develop dynamic, lower-criticality module increments

• Communicate proposed redirections

• Prepare for and execute acceptance tests, installation, operational evolution

• Analyze feedback on current version

• Reprioritize features

• Prepare strategy for next increment

• Use risk to determine content of artifacts

NSO and I3-led ProjectRisk Management Team

NISCM Strategy

6/25/03 ©USC-CSE 42

Risk Profiles Across Examples

Risk Items Risk Ratings Event Managers SupplyChain.com NISCM

Environmental Risks E1. Technology uncertainties E2. Many stakeholders E3. Complex system of systems Risks of using Agile Methods A1. Scalability — A2. Use of simple design — A3. Personnel turnover A4. Too-frequent releases A5. Not enough agile-capable people — Risks of using disciplined methods D1. Rapid change D2. Need for rapid results D3. Emergent requirements D4. Not enough discipline-capable people Risk rating scale: - minimal risk - moderate risk

- Serious but manageable risk - Very serious but manageable risk - Show-stopper risk

42©USC-CSE6/10/03

6/25/03 ©USC-CSE 43

Reflections on the Agile Manifesto

• The four value preferences are not exclusive-or’s – Processes and tools can empower individuals and

interactions– Documentation can help people understand complex

software – Contract negotiation can be an integral part of

customer collaboration– When changes are foreseeable, following a plan can

be the best way to respond to change

• Each pair of alternatives can reinforce each other – Can use risk to help find best balance

• People and their value propositions are top priority– Not just developers and customers, but end users,

maintainers, interoperators, general public, …

6/25/03 ©USC-CSE 44

Hybrid Agile/Plan-Driven Methods

• Early agility to determine emergent requirements– Agile user-interface prototyping

• Early planning to avoid surprises, point solutions– Plans for test data, drivers, oracles, tools, facilities– Architectures for assuring scalability, interoperability

• Early hybrid approach for complex COTS selection– Planned evaluation criteria, approach, priorities– Agile evaluations and iteration of approach

• Main development tailored to agile, plan-driven home grounds– Architect to modularize around sources of rapid change– Apply agile methods to rapidly changing lower-criticality

modules– Planned approach to multi-site relativity stable, higher-

criticality modules

6/25/03 ©USC-CSE 45

Large System Life Cycle Planning/Agility Framework

KPP’s:cost, schedule, performance, dependability, interoperability, …

Desired

FOC

DesiredIOC

Acceptable IOC

Desired FOC

Desired

Prioritized Capabilities

IOC Trade Space

Full Operational Capability(FOC) Architecture Support

AcceptableInitial Operational Capability (IOC)

Key

Per

form

ance

Par

amet

ers

(KP

P’s

)

6/25/03 ©USC-CSE 46

Conclusions

• Neither agile nor disciplined methods provide a silver bullet

• Agile and disciplined methods have home grounds where one clearly dominates the other

• Future trends are toward application developments that need both agility and discipline

• Some balanced methods are emerging• It is better to build your method up than to

tailor it down• Methods are important, but potential silver

bullets are more likely to be found in areas dealing with people, values, communications, and expectations management.