Integrating Beta Community Testing into Development using Gamification (EuroSTAR 2015)
Applying TPI An Unusual Case - EuroSTAR Conference · Agile processes: Scrum Small company Embedded...
Transcript of Applying TPI An Unusual Case - EuroSTAR Conference · Agile processes: Scrum Small company Embedded...
8/20/2009 © Endero Oyj 2005 1
Applying TPI – An
Unusual Case
Kari Kakkonen
Conformiq Software
© Endero Oyj 2005
2
THE INTELLIGENT ART OF INTEGRATION.
Applying TPI - an unusual case
Kari Kakkonen
Director, Testing and Consulting,
Endero Oyj
Dec 1st, 2005 at EuroSTAR
© Endero Oyj 2005
3
BACKGROUND ON PRESENTATION
Based on consulting work done during 2002-2005 and
consequent research paper finished on Feb 2005
Work done at Quality Assurance Consulting Services Unit of
Conformiq Software Ltd
The unit sold to Endero Ltd in June 2005, leaving Conformiq to
focus on model-based testing tools
Presenter personally part of all the assessments leading to
conclusions in this presentation, although in not all the
assessment part of Conformiq / Endero benchmark
© Endero Oyj 2005
4
KEY LEARNING POINTS
How to apply Test Process Improvement (TPI) to unusual
concepts without sacrificing the comparability to other TPI-scores
What kind of improvements to find in small companies (cases)
Multiple additions to TPI questionnaires for gaining more
information
© Endero Oyj 2005
5
ABSTRACT
TPI is often used with plan-driven processes in big companies.
TPI can also be used meaningfully in more agile companies, as
well as in maintenance testing.
One needs to be able to assess current testing capability in any
environment, identify the improvement needs, compare capability
and progress to own company and other companies, and get this
in concise form to get value out of money.
The session will describe three cases focusing on problematic
areas and their consequent improvements.
A comparison example to other companies will be shown to
pinpoint more opportunities of improvement.
© Endero Oyj 2005
6
NEEDS FOR ANY TEST PROCESS IMPROVEMENT
Objective is to identify
– ’ok’ –areas
– Critical improvement needs
– Improvement needs with biggest gain
– Smaller opportunities
Prerequisites
– Improvement needs to be continuous
– Related to actual work
– Benefit the overall process and business
Assumption
– Individual parts of testing process can be described and developed independently
– Dependencies can be named and also developed independently
© Endero Oyj 2005
7
LEVEL OF EFFORT
Time
Effort
Basic level,
continuous
improvement
Improvement
projects
Tailoring
© Endero Oyj 2005
8
TARGET LEVEL AND TIMETABLE
Time
Quality
Present level
Target level = ?
?Now
© Endero Oyj 2005
9
TPI IN SHORT
Test Process Improvement (TPI®) is a model for improving various testing processes
Developed and documented as a book by Tim Koomen and Martin Pol at Sogeti, the Netherlands
Systematic way to
– Assess software testing processes
– Find improvement opportunities
– Develop software testing step by step
20 key areas get a score based on fulfilling a number of checkpoints
Improvements relate to the score
Conformiq’s QA Consulting Services Unit (now part of Endero) cooperates with Tim Koomen
© Endero Oyj 2005
10
TPI MATURITY MATRIX EXAMPLE
A project
scores 0 to D
on all 20 key
areas
Key areas
with leftmost A
should be
developed
first
© Endero Oyj 2005
11
ASSUMPTIONS IN TPI
TPI-model seems to make some underlying assumptions, which
one can see by
– using TPI with an objective state of mind
– scrutinizing the checkpoints, dependencies and the book
These assumptions manifest themselves in
– Order and content of checkpoints
– Dependencies of key areas
– Improvement suggestions
© Endero Oyj 2005
12
ASSUMPTIONS
The assumptions seem to be that
– A certain order of improvements is always appropriate
– A software development organization always
– Is from medium to large size
– Has a separate, independent testing group
– Is very plan-driven
– Has fixed process phasing
– Has testing personnel with great level of autonomy
© Endero Oyj 2005
13
COMPARABILITY ISSUES
We find that the model can be applied and adjusted in many situations
One must take care of the comparability to other TPI-assessments
– The flexible points in the TPI model are when
– It’s only more or less information
– It doesn’t really matter
– Things that are needed to overcome the assumptions, but don’t alter the score, in the ”spirit of TPI”
– The corner stones that must be honored are especially
– Development order
– Big underlying assumptions such as independent test group
Comparability is needed to
– Utilize TPI fully – use it as intended as much as possible
– Be able to benchmark with other TPI-assessments
© Endero Oyj 2005
14
ASSESSMENT GROUND RULES
Scoring in key areas is only a secondary objective
Finding improvement needs is essential
Improvement goal has to be kept fresh in mind
Information is needed regardless of the way one gets it – ask any
questions necessary
© Endero Oyj 2005
15
SUGGESTIONS ABOUT POSSIBILITIES OF TPI
Take elements or ideas from other assessment models to get fuller picture of the test situation– E.g. if the object is to reduce the risk, one might also ask questions
inspired by Test Improvement Model (TIM)
Apply the context-driven analysis – adjust thinking from single checkpoint to thinking in terms of key area
level– based on the size of the assessed project and the process choices
Complete the big picture of the assessor by additional questions– Assessor experience– Other models, checklists, literature
Recognize both TPI-based and other realistic improvement suggestions
Prioritize the improvement suggestions and the key areas to be improved– By TPI– By additional perspectives
© Endero Oyj 2005
16
APPLYING MODELS
Choice of
models
Project
contextAssessor
expertise
Assessment
objectives
Assessment
with model
Model
suggestions
Applying
assessment
Applying
suggestions
Prioritized
suggestions
© Endero Oyj 2005
17
ADDITIONAL QUESTIONS TO BE USED WITH TPI
– What is included in your normal week of testing work?
– What are your roles in testing?
– How does testing relate to coding
– What is a typical project like, in terms of duration, size, type and resources
– How do you know what to test and when?
– What is the role and meaning of testing?
– How do you plan and prepare testing?
– What kind of build cycle do you have?
– Describe a typical test case?
– Describe skills and shortcomings of testing
© Endero Oyj 2005
18
GENERAL ABOUT CASES
Decision to use (also) something else than plan-driven approach
Small to medium sized companies or projects
– (quite typical in Finland)
Testing is (semi)integrated into the development team
Represent both software applications and embedded software
© Endero Oyj 2005
19
CASE ONE
Agile processes: Scrum
Small company
Embedded and application software
Plan and risk driven testing
Mostly developer with testing responsibilities, some full-time testers on project basis
High level of personal responsibility
(Data partially altered to anonymize)
© Endero Oyj 2005
20
SOME TPI ADJUSTMENTS MADE
Full-time requirement accepted for non-full-timers when no
conflicts of interests reported, enough time when testing
Generally approved daily build cycle replaces test manager
authorization for test environment installations
Meeting cycle is accepted as is, carefully assessing the objectives
of meetings against TPI checkpoints
© Endero Oyj 2005
21
SOME IMPROVEMENT SUGGESTIONS
Score: Half A, some B, Rest on initialized level
– (didn’t get any higher due to adjustments)
Reduce dependability on individuals by increasing (light)
documentation
Reserve (consultant) resources to automate routine testing
Include reviews into process officially (strategy, plans, checklists)
From verbal progress reporting into concise reports
Maintain structured approach on environment although decision
cycle during project is faster
© Endero Oyj 2005
22
CASE TWO
Small company
Embedded, critical software
Non-documented testing techniques
High personnel expertise
Developer has tester duties
Almost zero defects in production
(Data partially altered to anonymize)
© Endero Oyj 2005
23
SOME TPI ADJUSTMENTS MADE
Well-done just-in-time test case design accepted
Email and verbal reporting accepted, the level is essential, not the
means
Not suggesting to document everything
Trying to find out how to become more sustainable and
repeatable, adjusting heavily the order of TPI-based improvement
suggestions
© Endero Oyj 2005
24
SOME IMPROVEMENT SUGGESTIONS
Score: mostly initial level
– (didn’t get any higher due to adjustments)
Create a test strategy, prioritized
Document existing practices with simple checklists
Create a learning capability for new employees
© Endero Oyj 2005
25
CASE THREE
Several small projects in a fairly large company
– Average size 2-3 persons development effort in 1 month
– Separate acceptance testers part timing from their normal duties
Application software with large user base
Quick development cycle requires lighter organization and planning but
quite demanding level of software quality
TPI was appropriate to find improvements although some checkpoints
would automatically fail
Focus on acceptance testing
(Data partially altered to anonymize)
© Endero Oyj 2005
26
SOME TPI ADJUSTMENTS MADE
Full time test manager and tester requirements approved with
principle ”practically full time during project”
Any well-thought written record approved as document
– Level of planning assessed rigorously
Level and quality of communication assessed, not the means
Level of low-level testing assessed based on second-hand
information
© Endero Oyj 2005
27
SOME IMPROVEMENT SUGGESTIONS
Score: level A to B
– (didn’t get any higher due to adjustments)
– not possible to increase some key areas without violating TPI
– Given both prioritized suggestions and ”needed for TPI” suggestions
Testing environment needs a responsible support person
Test automation usage could increase from the level of using test management tools, as there is a centralized expert team in the company
Email communicated decisions should be stored to a central location
Checklists to be used with reviews
Keep the present moderate level of written planning
Test strategy thinking should be encouraged, not to be done only as a part of process
© Endero Oyj 2005
28
SCOPE OF IMPROVEMENT
The w
hole
com
pan
y
Time
Improvement
brainstorm
Ideas
proving
Piloting
ideas in a
project
Prepar
ation
Implementation
in project …x
Implementation
in project d…
Implementation
in project c
Implementation
in project b
Implementation
in project a
Piloting
ideas in
2nd project
Piloting
ideas in 3rd
projectAssessment:
1 project
represents
the company
© Endero Oyj 2005
29
BENCHMARK EXAMPLE
Project
should be
compared to
average
Key Areas Controlled Efficient Optimizing
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Test strategy A B C D
Life-cycle model A B
Moment of involvement A B C D
Estimating and planning A B
Test specification techniques A B
Static test techniques A B
Metrics A B C D
Test automation A B C
Test environment A B C
Office environment A
Commitment and motivation A B C
Test functions and training A B C
Scope of methodology A B C
Communications A B C
Reporting A B C D
Defect management A B C
Testware management A B C D
Test process management A B C
Evaluation A B
Low-level testing A B C
Min, max
Project
Average
© Endero Oyj 2005
30
BENCHMARK VALIDATION EXAMPLE
Key Areas Controlled Efficient Optimizing
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Test strategy A B C D
Life-cycle model A B
Moment of involvement A B C D
Estimating and planning A B
Test specification techniques A B
Static test techniques A B
Metrics A B C D
Test automation A B C
Test environment A B C
Office environment A
Commitment and motivation A B C
Test functions and training A B C
Scope of methodology A B C
Communications A B C
Reporting A B C D
Defect management A B C
Testware management A B C D
Test process management A B C
Evaluation A B
Low-level testing A B C
Min
Int’l avg
Conformiq avg
Int’l max
Conformiq max
Int’l avg
from Sogeti
survey
Conformiq/
Endero avg
from 30+
assess-
ments
© Endero Oyj 2005
31
CONCLUSION
Improvement suggestions can be found, regardless the size or
type of project or company
– TPI-score might not be superb – but that is beside the point
Improvements in the small projects tend to be
– in the same areas as in large projects, but to be achieved with
different means or level of intensity
– prioritized differently
TPI-assessments can be enhanced with
– Context-driven touch to the company
– Ideas from other models
– Additional questions and checklists
© Endero Oyj 2005
32
CONTACT INFORMATION
www.endero.com
Endero Ltd
Kiviaidankatu 2 F
00210 Helsinki
Finland
© Endero Oyj 2005
33
ENDERO IN SHORTThe intelligent art of integration.
Software Services Partner for IT-departments and software
companies
Services include software specialist and testing services
Primary market: Finland
Founded: 1994
Traded at Helsinki Stock Exchange since 1999
59 persons (in the end of 2004)
Turnover 5,5 M€ (2004)
Offices: Helsinki