In the international bestseller 'Thinking, Fast and Slow', Daniel Kahneman explains how we as human beings think and reason, and perhaps surprisingly how our thought processes are often fundamentally flawed and biased. This talk explores the ideas presented in the book in the context of professional software development. As software developers we all like to think that we are highly logical, and make only rational choices, but after reading the book I'm not so sure. Here I'll share my thinking on thinking. Topics that will be discussed include; the 'Availability Heuristic', which can lead developers to choose the 'latest and greatest' technology without proper evaluation; 'Optimistic Bias' which can blind architects from the 'unknown unknowns' within a project; and more!
Devternity 2016 "Thinking Fast and Slow with Software Development"
1. Thinking, Fast and Slow…
With Software Development
Daniel Bryant
Chief Scientist, OpenCredo
daniel.bryant@opencredo.com
@danielbryantuk
2. Think More Deliberately
• Our decision making can be flawed…
• Apply process and models (as appropriate)
• Collaborate more (and better)
• Plan, do, check, act…
03/12/2016 @danielbryantuk
3. @danielbryantuk
• London Java Community Associate
• InfoQ Editor, DZone MVB, O’Reilly
• Conference regular…
03/12/2016 @danielbryantuk
• Chief Scientist at OpenCredo
• Technical/digital transformation
• Java, Golang, CI/CD, DevOps
• Microservices, cloud and containers
5. Our Thinking: A Tale of Two Systems…
System 1:
fast, instinctive, emotional, subconscious
System 2:
slower, deliberate, reasoning, conscious
03/12/2016 @danielbryantuk
12. Our Thinking: A Tale of Two Systems…
System 1:
fast, instinctive, emotional, subconscious
Rapid, associative, and has systemic errors
System 2:
slower, deliberate, reasoning, conscious
Lazy, and causal (not statistical)
03/12/2016 @danielbryantuk
14. Availability Heuristic
“If something can be recalled,
it must be important”
‘Hipster-itis’
e.g. the ‘best’ architectural style
03/12/2016 @danielbryantuk
15. Microservices
• The current flavour of the month!
– Great ideas, but probably over-hyped
• Frameworks and products emerging
– Virtuous (vicious?) circle
– “The bandwagon effect”
– Have we found the silver bullet?…
03/12/2016 @danielbryantuk
16. Whatever is Available
When all you have is a hammer…
…everything looks like a nail
03/12/2016 @danielbryantuk
17. When all you have is a SOA…
…everything looks like a service
03/12/2016 @danielbryantuk
18. When all you have is a Jenkins…
…everything looks like a Jenkins Job
03/12/2016 @danielbryantuk
19. When all you have is a Java Spring Framework…
…everything looks like an AbstractSingletonProxyFactoryBean
03/12/2016 @danielbryantuk
22. Evaluation
“I will postpone using this shiny new framework
until my peers have validated the proposed
benefits with rigorous scientific experiments”
- Said by no programmer
…ever
03/12/2016 @danielbryantuk
24. Optimistic Bias
“People tend to be overconfident,
believing that they have substantial
control in their lives”
I know what our customers want…
…how could I possibly be wrong?
03/12/2016 @danielbryantuk
25. Four Factors of Optimistic Bias
• Desired end state
– Self-enhancement, perceived control
• Cognitive mechanisms
– Representativeness heuristic, singular target focus
• Information about self vs target
• Overall mood
03/12/2016 @danielbryantuk
26. Optimism: Think Professionally
• Define (and share) clear goals
• Build, measure, learn…
• Remove uncertainty early (bit.ly/1mAb6o4)
– “Patterns of Effective Delivery” by Dan North
• Software is inherently collaborative…
03/12/2016 @danielbryantuk
30. “Developer-on-call”
An occasional spike to the head
is a good thing...
...metaphorically speaking
• You build it, you run it
– Shared responsibility
– Communication
03/12/2016 @danielbryantuk
33. Planning Fallacy
“A phenomenon in which predictions about
how much time will be needed to complete a
future task display an optimistic bias. ”
Was your last project completed on time?
…and on budget?
03/12/2016 @danielbryantuk
34. IT Track Record…
• Sainsbury’s Supply Chain Management System
– $526m bit.ly/160SnAj
• NHS patient record system
– £10bn bit.ly/XBzFuV
• HealthCare.gov
– onforb.es/1k7egyb
03/12/2016 @danielbryantuk
35. Most Common Factors for Failure
• Unrealistic or unarticulated project goals
• Inaccurate estimates of needed resources
• Badly defined system requirements
• Poor reporting of the project's status
• Unmanaged risks
• Poor communication among customers, developers, and users
• Use of immature technology
• Inability to handle the project's complexity
• Sloppy development practices
• Poor project management
• Stakeholder politics
• Commercial pressures
Source spectrum.ieee.org/computing/software/why-software-fails
03/12/2016 @danielbryantuk
36. At the Start: Feedback & Federation
• Determine goal or hypothesis
– Define ‘metrics of success’
• Plan, do, check (showcase), act
• Divide and conquer
– SOA, microservices, modules… (systems thinking)
– Squads, guilds, chapters
03/12/2016 @danielbryantuk
43. Sunk Cost Fallacy
“Any past cost that has already been
paid and cannot be recovered should not
figure into the decision making process. ”
When did you last remove a framework?
…or a library?
03/12/2016 @danielbryantuk
44. Why Are We Reluctant?
• We don’t like being wrong…
• Existing effort appears wasted
– Endowment effect
• Loss aversion
– Twice as powerful, psychologically, as gains?
03/12/2016 @danielbryantuk
45. Try not to ‘Sink Costs’
03/12/2016 @danielbryantuk
46. Retrospect Regularly
• Did we make the right choice?
• When was the ‘last responsible moment’?
• What can we learn?
• How can we get better?
03/12/2016 @danielbryantuk
47. Anchoring Bias
“Common tendency to rely too heavily
on the first piece of information offered
when making decisions. ”
How does your manager ask for estimates?
…is it an unbiased question?
03/12/2016 @danielbryantuk
49. Anchoring: Learn to “Upwardly Manage”
• Learn to say no…
– Provide explanations and alternatives
• Make sure goals/user stories are well-defined
– Collaboration
– Feedback
• Apply PERT estimations (bit.ly/1mGzuoe)
03/12/2016 @danielbryantuk
50. A Little More #NoEstimates
03/12/2016 @danielbryantuk
ronjeffries.com/xprog/articles/the-noestimates-movement/
51. Read Your Way to Tech Leadership (?)
03/12/2016 @danielbryantuk
53. System 1 is Awesome!
• It saves us from many dangers
• Very valuable in many situations
– ‘gut feel’, instincts, intuition
– Expertise plays a role
– Can be used for good
• But, I see too much System 1 in IT…
03/12/2016 @danielbryantuk
54. We All Make Mistakes…
03/12/2016 @danielbryantuk
56. “You will be the same person in five years as
you are today except for the people you meet
and the books you read.”
- Charlie “Tremendous” Jones
(bit.ly/1LAdQkv)
03/12/2016 @danielbryantuk
59. Summary
• Apply process and models (as appropriate)
– Listen to system 1, but engage system 2…
• Collaborate more (and better)
• Learn, do, retrospect, (teach,) repeat
- “Think” more deliberately -
03/12/2016 @danielbryantuk
60. Thanks for Listening!
Comments and feedback are welcomed…
daniel.bryant@opencredo.com
@danielbryantuk
www.infoq.com/author/Daniel-Bryant
www.parleys.com/author/daniel-bryant
www.muservicesweekly.com
03/12/2016 @danielbryantuk
Notes de l'éditeur
Many explanations for the optimistic bias come from the goals that people want and outcomes they wish to see.[1] People tend to view their risks as less than others because they believe that this is what other people want to see. These explanations include self-enhancement, self-presentation, and perceived control.
Self-enhancement suggests that optimistic predictions are satisfying and that it feels good to think that positive events will happen.
People tend to focus on finding information that supports what they want to see happen, rather than what will happen to them
Representative heuristic: individuals tend to think in stereotypical categories rather than about their actual targets when making comparisons. The estimates of likelihood associated with the optimistic bias are based on how closely an event matches a person's overall idea of the specific event
Individuals know a lot more about themselves than they do about others.[1] Because information about others is less available, information about the self versus others leads people to make specific conclusions about their own risk, but results in them having a harder time making conclusions about the risks of others. This leads to differences in judgments and conclusions about self-risks compared to the risks of others, leading to larger gaps in the optimistic bias
In behavioral economics, the endowment effect (also known as divestiture aversion) is the hypothesis that people ascribe more value to things merely because they own them
(1990 where participants were given a mug and then offered the chance to sell it or trade it for an equally priced alternative good (pens). Kahneman et al. (1990)[2] found that the amount participants required as compensation for the mug once their ownership of the mug had been established ("willingness to accept") was approximately twice as high as the amount they were willing to pay to acquire the mug ("willingness to pay")
Loss aversion implies that one who loses $100 will lose more satisfaction than another person will gain satisfaction from a $100 windfall.
delay commitment until the last responsible moment, that is, the moment at which failing to make a decision eliminates an important alternative. If commitments are delayed beyond the last responsible moment, then decisions are made by default, which is generally not a good approach to making decisions.
The key is to make decisions as late as you can responsibly wait because that is the point at which you have the most information on which to base the decision.
Program Evaluation and Review Technique
Three-point estimation technique