Collaboration Without Chaos - STP Spring 2013

  • View
    4.492

  • Download
    0

Embed Size (px)

DESCRIPTION

Sometimes software testers overvalue the adherence to the collective wisdom embodied in organizational processes and the mechanical execution of tasks. Overly directive procedures work—to a point—projecting an impression of firm, clear control. But do they generate test results that are valuable to our stakeholders? Is there a way to orchestrate everyone’s creative contributions without inviting disorganized confusion? Is there a model that leverages the knowledge and creativity of the people doing the work, yet exerts reliable control in a non-directive way? Griffin Jones shares just such a model, describing its prescriptive versus discretionary parts and its dynamic and adaptive nature. Task activities are classified into types and control preferences. Griffin explores archetypes of control and their associated underlying values. Leave with an understanding of how you can leverage the wisdom and creativity of your people to make your testing more valuable and actionable. Note: There is a lot of technical detail, and the animation being flattened out of the slides makes it harder to understand than when I present it. You also don't have have stories, explanation, or the ability to answer your questions.

Text of Collaboration Without Chaos - STP Spring 2013

Slide 1

1 www.CongruentCompliance.com

COLLABORATION WITHOUT

April 2013

SESSION 702 Collaboration Without Chaos

Griffin Jones, Consultant, Congruent Compliance

2 www.CongruentCompliance.com

COLLABORATION WITHOUT

April 2013

Introduction

About me Questions

What does control look, sound, feel like? What does chaos look, sound, feel like? What collaboration look, sound, feel like?

Why I care Regulated and risky products

I have to explain my approach to skeptics My early formative career experiences

Nuclear Industry and KIMS

April 2013 3 www.CongruentCompliance.com

Qualifying Human Experts

How many faces did you see?

Qualifying Human Experts

Count the number of faces in each of the following

pictures

Fast Example Exercise

April 2013 www.CongruentCompliance.com 4

Motivation to Create

Was asked, How do you delegate control? Dissatisfied w/some answers I hear/see from others

As little as possible: One big brain, controlling people like machines/puppets

Seems to be about people managing or coping w/their anxiety My ANS: Competence, Authority, Responsibility, Experience

I noticed I had a strong emotional reaction to the Q&A I prefer

Collaborating with thinking sapient people who test Giving tactical control of check activities to thinking people

They choose how/when to use tools like automation and scripted procedures, to do what the tools do best no more.

Tools are made to serve people, not Just because I could micro-manage to deal with my anxiety,

April 2013 5 www.CongruentCompliance.com

this Presentation

Realized that I was being intentional But some archeology of my process was required

Where did all the pieces come from? How do they all fit together?

Benefits of a Grand Unified Model Adaptive Discretionary Control Model Creates (for me) a framework for conversation,

exploration, and formal explanation By sharing this, I hope it is helpful to you

Stimulates you to explore your own thoughts Creates space to allow thinking collaborative testing

April 2013 6 www.CongruentCompliance.com

Historical Example: The Flying Problem

Powered flight research prior to 1900 Focus was on Control via

Inherent Stability versus Dynamic Stability

Why did they make that control choice? How is the situation similar to testing software?

April 2013 7 www.CongruentCompliance.com

Langley Aerodrome

Wright Flyer I 2 3

Outline: Collaboration Without Chaos Adaptive Discretionary Control Model

Fit the pieces together from these models: Feedback Controllers

and the Controllers Internal Model Administrative Controls

Prescriptive versus Discretionary Controls Control Choices

Organizational and Activity Types. How to choose. OODA Model (Observe/Orient/Decide/Act)

The Orient process Values Archetypes

Example Values to Orient on Adaptive Team Model

The three ways to adapt Summary / Thought Experiment / References / Questions

April 2013 8 www.CongruentCompliance.com

Spoiler Alert! Two Big Points,

A. Chaos is the manifestation of inconsistency. It is a sign that the controller is failing

B. To get Collaboration, you need to value it and be consistent You cant have collaboration using primarily

prescriptive procedures even if your values are aligned to collaborate

You cant have collaboration using discretionary procedures

if you values are primarily aligned against collaboration

April 2013 9 www.CongruentCompliance.com

Values Archetypes

(Inspired by Italy Talgam)

Adaptive Team Model ATM

(Entin, Serfaty, Deckert)

Feedback Controllers and the

Controllers Internal Model

(Jerry Weinberg)

How to Make Control Choices

Organization and

Activity Types

(Reasons & Perrow)

Administrative Controls

Prescriptive versus

Discretionary

(James Reasons)

OODA Model

Observe Orient Decide

Act

(John Boyd)

Finding the Path through these Models

April 2013 10 www.CongruentCompliance.com

4 - 10

Adaptive Discretionary Control Model

Feedback Controllers

April 2013 11 www.CongruentCompliance.com

4

Feedback Controllers - Overview

The system of control must have an image of the desired state (D) the ability to observe the actual state (A) (Testing Info & Other Outputs) The ability to compare state A and D for differences The ability to act on the system to bring A closer to D. (Resources & Req.)

Control is exercised through Resources and Requirements

April 2013 12 www.CongruentCompliance.com

Controller

Testing Information

Other Outputs

Req.

Testing Resources

Randomness

Req.

Resources

Randomness

Req.

Testing Information

Other Outputs

Controller

Feedback Controllers Model

The system of control must have an image of the desired state (D) the ability to observe the actual state (A)

(Testing Info & Other Outputs) the ability to compare state A and D for differences

April 2013 13 www.CongruentCompliance.com

Testing Information

Other Outputs Testing

Testing Information

Other Outputs

Desired State (D) or

Model Controller

Actual State (A) or

Information

?

11

Administrative Controls

April 2013 14 www.CongruentCompliance.com

5

Intermittent additions of organizational wisdom to Principles, Rules, Procedures as a result of Incidents

and Non-Conformance.

Frequent comparisons of output measures w/org. objectives. Deviant performance corrected.

Frequent comparisons of output measures w/org. objectives. Deviant performance corrected. Congruent performance stored

as rules & procedures.

Administrative Controls

April 2013 15 www.CongruentCompliance.com

OUTPUT

Feedforward / Prescriptive Feedback / Discretionary

Organizational Standards, Objectives

Principles, Rules, Procedures

Human Performance

Training, Experience, Using Tools

Process being

Controlled

Output Measures

X

Zero Deviation

Incidents, Non-Conformance

Integrating Prescriptive/Discretionary Controls with Resources/Requirements

April 2013 16 www.CongruentCompliance.com

Testing Testing Information

Other Outputs

Controller ?

Prescriptive and Discretionary

Administrative Controls Req.

Resources

Randomness

April 2013 17 www.CongruentCompliance.com

How to Make a Control Choice

6

Rules and procedures not applicable. Task performance at

the discretion of the individual.

Requires a mixture of prescriptive control by rules and procedures and discretionary performance by the individual.

Requires mixture of prescriptive and discretionary performance control.

Pre-programmed prescriptive process control possible by rules and procedures.

April 2013 18 www.CongruentCompliance.com

Determine Control Choice based on Exception / Solution Characteristics of Activity

FEW EXCEPTION CASES

MANY EXCEPTION CASES

EASY SEARCH

FOR SOLUTIONS

HARD SEARCH

FOR SOLUTIONS

Tasks routine, repetitive, well-structured and predictable.

Tasks non-routine, but the many exceptional cases are relatively simple to analyze.

Work routine, but problems are sometimes vague and poorly conventionalized.

Tasks non-routine, poorly structured and unpredictable.

April 2013 19 www.CongruentCompliance.com

Control Choices for Varieties of Organizations

FEW EXCEPTION CASES

MANY EXCEPTION CASES

EASY SEARCH

FOR SOLUTIONS

HARD SEARCH

FOR SOLUTIONS

Railways Postal Services Construction Traditional Banking Road Haulage

Nuclear Power Plants Chemical Process Plants Modern Aircraft Advanced Manufacturing Anesthesia

Architecture Maintenance and Repair Oil Exploration Police Work Scientific Research

Modern Military Operations Investment Banking Macro-economics Crisis Management

Project Management

Recovering from Design Basis Accidents / Issues

Production Lines

R&D Organizations

PRESCRIPTIVE MIXTURE

MIXTURE DISCRETION

Integrating Control Choice with Administrative Controls

Analyze the organization and task activities to make Control Choices Exception Cases Search for Solutions

April 2013 20 www.CongruentCompliance.com

Testing Testing Information

Other Outputs

Controller ?

Prescriptive and Discretionary

Administrative Controls Req.

Resources

Randomness

Task: Exception Cases Search for Solutions

The OODA Model

April 2013 21 www.CongruentCompliance.com

7

The OODA Model

April 2013 22 www.CongruentCompliance.com

OBSERVATIONS