Developing Developers

1
Developing Developers
2
Context

• Your company has been repeated criticized by
  customers for producing poor products.

3

• For instance, the most recent product, resulting
  from 45-million lines of code, was released on
  the same day as 18 megabytes of patches - bug
  fixes and enhancements.

4

• The fixes were presented as two patches. One of
  the patches did not work.

5
Response

• The CEO called on employees to make reliable and
  secure computing their priority, and demanded a
  reduction in the defects in the products they
  produce.

6

• As part of corporate strategy, all code writing
  was halted for two months so developers could
  undergo training on reliability and security. The
  training was provided by the company and prepared
  by their Programmer Productivity Center.

7
PPRC Goal
The goal of the Programmer Productivity Research
Center (PPRC) at Microsoft Research is to
investigate radical approaches to help Microsoft
deliver high-performing and reliable software.
8
Perspective
Modern software programs are massive, complex
structures. As a result, software
development is increasingly
labor-intensive. We believe the answer lies in
good tools and processes. Good tools enable
process changes and allow a large number of
developers and testers to effectively work
together.
9
Critique

• How would your evaluate the CEOs approach to
  this situation?
• Do you feel the PPRC provides a reasoned
  perspective on the issues?

10
Taylorism Revisited

• Objectification of Work
• Standardization of tools
• Standardization of work methods
• External authority specifies procedures

11
Productivity

• Tool Focus
• A fool with a tool is still a fool!

12
Basis
New hires do not know how to communicate
effectively, they have insufficient experience
and preparation for working as part of a team,
they lack the ability to manage their individual
work efficiently and productively, and they do
not understand or appreciate organizational
structures or business practices. Hilburn, 1997
13
Kinds of Knowledge (1)

• declarative knowledge (knowledge that)
• refers to the kind of knowledge typically learned
  from textbooks--facts and concepts

14
Kinds of Knowledge (2)

• procedural knowledge (how to knowledge)
• refers to being able to do something, be it
  writing code, proceeding through analysis and
  design

15
Manifestations of Expertise

• experts appear to have much more abstract
  representations
• experts are able to apply what they know to a
  broader range of problems
• experts evidently have a script that guides the
  conduct of their work
• experts have better developed metacognitive
  processes

16
Kinds of Knowledge (3)

• metacognitive knowledge
• refers to a persons skill at planning strategy,
  monitoring process and progress, changing what
  one is doing when appropriate, and reflecting on
  the process so that one can discover ways to
  improve.

17
Metacognition

• People with good metacognitive skills are better
  at problem solving.
• People with good metacognitive skills perform
  better on ill-structured tasks.
• People with good metacognitive skills perform
  better in far-transfer tasks.

18
The Road to Expertise

• it takes a long time
• it takes intense, focused effort
• it can not be taught directly

19
Caveat!

• Practice alone is sufficient to develop a
  certain level of competence, but, for most
  people, will result in an asymptotic level below
  that of true excellence.

20
Single Loop Learning
From Peter Senge, et. al (2000) Schools that
Learn.
21
Double Loop Learning
From Peter Senge, et. al (2000) Schools that
Learn.
22
Approach Knowledge Distillation

• Sterilized view
• Focus on technical details
• Primacy of knowing-that
• Use of knowledge-telling strategies

23
Approach Identify and Nurture

• Early talent does not predict later success.
• Ability measures not predictive of job
  performance.
• Measures of initial ability predict entry level
  performance, but do not predict long-term job
  success.

24
Approach Practice, Practice, Practice

• Focus tends to be on products
• Unsure what is really learned
• Experience only a weak predictor
• Deliberate practice

25
Approach Situated Cognition

• Project Orientation
• Groups
• Technical and Non-Technical Issues
• Occurs late in curriculum

26
Constructivist

• Knowledge is an individual(social) construction.
  The learner interprets and constructs a reality
  based on his experiences and interactions with
  his environment.
• Learning is an ACTIVE, not passive, process.

27
Mental Models

• Clearly the constructivists posit the existence
  of cognitive structures within the individual.
  These mental models provide the individual with
  guidance in terms of dealing with tasks and
  situations.

28
Transfer

• Near
• Far
• Type of task?
• Error Recovery
• Problem Solving

29
Cognitive Apprenticeship
30
Things to Learn
31
Conclusion

• Nature of the system
• Nature of the environment
• Nature of the learner