Post on 29-May-2020
© Alistair Cockburn 2015
“Heart of Agile” meets “Risk Reduction 2.0 (Disciplined Learning)”
Dr. Alistair Cockburn http://Alistair.Cockburn.us
© Alistair Cockburn 2015
2
People ...
To an interpreter unforgiving of error
Making decisions every decision has economic consequences and resources are limited.
… they don’t understand which keeps changing
Deciding
… they don’t understand which keeps changing … they don’t understand which keeps changing
© Alistair Cockburn 2015
3
The Agile Manifesto invites wimpy-ness
"… Individuals and interactions over processes & tools…" (Yayy!! I don't have to follow those stupid processes any more!)
"… Working software over comprehensive documentation…"
(W00t!! Dump the documentation! I LOVE this agile stuff!) "… Customer collaboration over contract negotiations…"
(I'm done when I'm done and I never have to say when!) "… Responding to change over following a plan…"
(No plans! No project managers! No architects! )
Where do I sign up?
© Alistair Cockburn 2015
4 People develop skills in 4 stages
© Alistair Cockburn 2015
5
The Heart of Agile simplifies again
Collaborate
Deliver
Reflect
Improve
© Alistair Cockburn 2015
6
The Heart expands into Details
Learning Income
Collaboration
Trust
Experiment Change
Examine
Introspect
Collaborate
Deliver
Reflect
Improve
Business Social Technical
Cost Queues Early Revenue
Goals
Results
Include Emotions
Focus Forwards
Emotional Safety
Concretely (Solutions Focus)
Limit Changes
Aggressively
Allow failure
Let someone else do it
Step forward
Listen
© Alistair Cockburn 2015
7
Craft Professions
skills in a medium
Cooperative Game communication teamwork strategies
Knowledge Acquisition pay-to-learn trim-the-tail
Flow Management unvalidated decisions
= inventory, lean processes
Self-Awareness (Personalities) the people, strategies, techniques
actually used
Advanced Agile Development
origami courtesy of Gery Derbier
© Alistair Cockburn 2015
8
Use Lean Processes
for Flow Management
© Alistair Cockburn 2015
9
(Think about how many lines
of code testers get
handed at one moment!)
In software development, people wait on other people to make decisions
I wish they’d decide what style they want!
Users & Sponsors
Testers Business Analysts
I wish they’d decide what
functions they want!
A decision made by the user/sponsor to be used by the
UI designer I wish they’d decide
on the UI design already!
I wish they’d think a bit more carefully
before coding!
Programmers
Every line of code is a decision!
UI Designers
© Alistair Cockburn 2015
10
Users & Sponsors
Testers
Business Analysts
Programmers
UI Designers
Software = manufacturing if Inventory = Decisions!
© Alistair Cockburn 2015
11
A team’s optimal process and best strategies depend on its decision dependency network.
Users & Sponsors
absent
UI Designers
Business Analysts
some Programmers
Database designers
Tester
Enough Users & Sponsors
Many Business Analysts
Not enough Programmers Some
Testers
Enough Users & Sponsors
UI Designer
A few Business Analysts
Lots of Programmers
Single, solitary Database designer
© Alistair Cockburn 2015
12
Large-batch transfers (of decisions, in our case ) makes a system’s behavior harder to predict.
time
people
Vertical lines = queue length at any time
Horizontal lines = wait time in the queue
time
People arriving
People being served
people
© Alistair Cockburn 2015
13
Queuing theory shows that feed systems like these run more smoothly with small transfers
time
Vertical lines = queue length at any time
Horizontal lines = wait time in the queue
People arriving
People being served
time
people people
© Alistair Cockburn 2015
14
This applies when you choose to hand testers 1,000 lines of code to test, or 100 lines of code.
Hint: Prefer 100
time
Testers 100 lines of code show up for testing
Total # of design decisions being validated
Decisions being validated (or broken)
time
Testers 1,000 lines of code show up for testing
How long this line of code / screen / use case /
decision sits in the testing queue
How long this line of code / screen / use case /
decision sits in the testing queue
Decisions arriving
© Alistair Cockburn 2015
15
Users & Sponsors
UI Designers
Programmers Testers
Business Analysts
Convert jams
Users & Sponsors
UI Designers
Programmers Testers
Business Analysts
to continuous flow.
Reduce decisions in motion & balance the flow
© Alistair Cockburn 2015
16
Design is
Knowledge Acquisition
© Alistair Cockburn 2015
17
Delivers nearly no knowledge (or risk reduction)
Big-Bang Design is a late-learning strategy
time
Knowledge comes at the “moment of truth”: final integration.
cost
Growth of knowledge with big-bang integration
© Alistair Cockburn 2015
18
Development sequence indifferent (with respect to knowledge)
Delivers knowledge (risk reduction)
We can pay to learn early in the project
time
cost
Growth of knowledge with early, continuous integration
© Alistair Cockburn 2015
19
Develop for business value once risks are down
time
Knowledge growing (risk reduction)
cost
Business value growing
Reduce risks: - Business - Social - Technical - Cost / Schedule
© Alistair Cockburn 2015
20
Value
Knowledge / risk reduction
Three project stages: Risk, Value, Tail
time
Tail
risk value tail Item1 .. … ... item 2 . … .. item 3 .. … . item 4 .. .. . :
© Alistair Cockburn 2015
21
Trim to deliver on-time (or early)
Delay to get more or better
Payoff = Trim-the-Tail: Deliver by value or date
© Alistair Cockburn 2015
22
Craft Professions
skills in a medium
Cooperative Game communication teamwork strategies
Knowledge Acquisition pay-to-learn trim-the-tail
Flow Management unvalidated decisions
= inventory, lean processes
Self-Awareness (Personalities) the people, strategies, techniques
actually used
Designing in Teams
origami courtesy of Gery Derbier
© Alistair Cockburn 2015
23
Postscript: time to deliver a project
time = log( D (b..p) ) * N (u..p) * log( N (p) ) time to deliver a project =
distance between business owner & programmer (log)
* number of people between user & programmer
* number of programmers (log)
© Alistair Cockburn 2015
24
Heart of Agile
Collaborate
Deliver
Reflect
Improve