Hybriding CMMI and Requirement Engineering Maturity & Capability Models - Applying the LEGO Approach for Improving Estimates

ICSOFT 2012 – 7° Int. Conference on Sw Paradigm Trends
Rome (Italy), 24-27 July, 2012

Hybriding CMMI and Requirement Applying the LEGO Approach
Engineering Maturity & Capability for Improving Estimates
Models

Luigi Buglione
Jean C.R. Hauck
Christiane Gresse von Wangenheim
Fergal McCaffery

www.eng.it

DKIT At a glance

Dundalk Institute of
Technology is a 90 acre
campus situated between
Dublin and Belfast (each
approximately 50 miles
away).

The Institute consists of 4
Schools:

1. Business & Humanities
2. Informatics & Creative
Arts
3. Engineering
The Regulated Software Research Group which is 4. Health & Science
part of LERO (the Irish Software Engineering
Research Centre) is part of the School of Informatics
& Creative Media

www.eng.it

Engineering At a glance

_ The first Italian ICT player
_ more than 730 M/€ revenues Research and PA & HC Finance Industry TELCO Utilities
Development
_ 1000 clients
_ 6,300 IT specialists System Int. &
Consultancy
% 46 70 54 80 80

Outsourcing % 35 10 27 10

Software
% 19 20 19 10 20

ERP IT Security ECM

Plant Management
Managed Operations Broadband & Media
System

www.eng.it

www.eng.it

UFSC At a glance

Federal University of Santa Catarina
Florianópolis/Brazil [http://www.ufsc.br]

• 48 Master courses
• 38 PhD courses
• 62 Undergraduate courses
• EUR 268 million annual budget
• 25,737 Undergraduate students
• 8,543 Graduate students
• 34,280 Students
[http://www.youtube.com/watch?v=V6E1Z5DEuvk]

INCoD is to be a reference institute for excellence in research, validation
and dissemination to support digital convergence. [http://www.incod.ufsc.br]

The Software Quality Group focuses on scientific research, development and
transfer of SE models, methods and tools. [http://www.gqs.ufsc.br]

www.eng.it

Hybriding CMMI... Goals of the presentation

 G1. Discuss the relevance of the Req. Engineering (RE)
process as the main input for the Estimation process
 G2. Present the state-of-the-art of RE Maturity &
Capability Models (MCMs)
 G3. Show how – using the LEGO approach - the
‘hybridization’ among 2+ RE MCMs practices can improve
the Estimation process

ICSOFT 2012 – Rome (Italy), 24-27 July, 2012
5
© 2012 Buglione, Hauck, Gresse von Wangenheim, McCaffery

Hybriding CMMI... Agenda

• Introduction
– Being…Glocal!
– The Estimation Process
• Requirement Engineering: Some Maturity & Capability
Models (MCM)
– The MM-Mania
– RE models and LEGO
– Some RE maturity & capability models (MCM)
• Experiencing LEGO to Requirement Engineering
– The LEGO Approach
– Applying LEGO to RE
– EoI: Elements of Interest
– Suggestions for Improvement
• Conclusions & Next Steps
• Q&A

6

Introduction Being...Glocal!

• ‘Glocal’ = think globally, act local (Swyngedouw, 1997)

Western world –
People act
Eastern world:
People think (a
bit more) before
acting

• Q: …thus, which could be the best practices for improving Estimation?

7

Introduction The Estimation Process (Lifecycle view)

Anything starts here...

Some known issues for wrong estimates:
 Missing/incomplete requirements
 More attention paid to formalize FUR than NFR
 Gather too high-level requirement (no granularity)
 ‘Scope creep’ phenomenon
 Not optimal stakeholder engagement
…

• Q: can we find solutions/tips from Maturity & Capability Models ( MCMs)?

8

Introduction The Estimation Process (CMMI-DEV, ML2)

MA – Measurement & Analysis PP – Project Planning

SG1
Establish Planning Data
Estimates
Measurement
Data
SG2
Develop a Project Plans
Project Plan

An agreed-to set
of requirements SG3 Obtain
Committment
to the Plan

REQM – Requirement Mgmt PMC – Project Monitoring & Control

9


• Introduction
– Being…Glocal!
Models (MCM)
– The MM-Mania
• Q&A

10

RE – Some MCMs ‘MM-mania’: UFSC/DKIT research

Source: Gresse von Wangenheim C., Hauck J.C.R., Zoucas A. Salviano C.F., McCaffery F., Shull F., Creating Software Process Capability/Maturity
Source
Models. IEEE Software, vol. 27 no. 4, pages 92 -94, July/August 2010.

11

RE – Some MCMs ‘MM-mania’: SEMQ website

URL: www.semq.eu/leng/proimpsw.htm
URL

12

RE – Some MCMs MCM Repository - www.gqs.ufsc.br/mcm

• Web-based repository
 Java v5.0 on a web platform
 Tomcat v6.0 application server
 MySQL Community Server v5.5

Source: Gresse von Wangenheim C., Hauck J.C.R., Buglione L., Mc Caffery F., Cardoso Lacerda T., Viera da Cruz R.F., Building a Maturity & Capability
Source
Model Repository, Proceedings of PROFES 2011, 12th International Conference Product Focused Software Development and Process Improvement,
Torre Canne - Bari (Italy), June 20-22 2011, Second Proceedings: Short Papers, Doctoral Symposium and Workshops, ACM, ISBN 978-1-4503-0783-3,
pp. 2-5

13

RE – Some MCMs Classifying MCMs by Dimension

Source: Buglione L., An Ecological View on Process Improvement: Some Thoughts for
Improving Process Appraisals, 4WCSQ, 4th World Congress on Software Quality,
Washington D.C. (USA), 15-18 September 2008
• Horizontal: MMs going through the whole supply chain
 SwEng: ISO/IEC 15504 (SPICE), CMMI, FAA i-CMM, …
• Vertical: MMs focusing on a single perspective/group of processes
 Test Mgmt: TMM, TPI, …
 Project Mgmt: PM-MM, OPM3, …
 Requirement Mgmt: ....
• Diagonal: MMs focused on Organizational/Support processes
 People CMM, TSP, PSP, …

14

RE – Some MCMs Choosing RE MCMs - Results

Model/ Architect-
Repr. Type ML (#) Comments/Notes
Framework Type
IBM RMM Staged 6 [0-5] Level-based ---
(Heurmann, 2003)
(Sehlhorst, 2007)
IAG RMM (IAG, Staged 6 [0-5] Matrix-based 6 dimensions (process, practices & techniques,
2009) deliverables, technology, organization, staff
competency)
PRTM CRMM Staged 4 [0-3] Level-based ---
(Hepner, 2006)
BTH REPM Gorschek Continuous Process-based 7 processes
et al, 2002) Variable number of sub-process areas per process
(Gorschek, 2011)
REAIMS Process Staged 3 [1-3] Process-based 8 process areas and 66 practices (basic, intermediate,
MM (Sommerville, advanced)
2005)
R-CMM Staged 5 [1-5] Process-based ‘Processes’ = Practices (e.g. 20 ‘processes’ at ML2) -
(Beecham et al. Adaptation of GQM for deriving practices
2005) (Beecham et
al., 2003)
R-CMMi (Solemon et Staged 5 [1-5] Process-based ‘Processes’ = Practices (e.g. 20 ‘processes’ at ML2) -
al., 2009) Adaptation of GQM for deriving practices using the
CMMI process architecture

15


• Introduction
– Being…Glocal!
Models (MCM)
– The MM-Mania
• Q&A

16

Experiencing LEGO... The LEGO Approach

1. MCM Repository 2. Process
Architecture

3. Mappings & 4. Appraisal Method
Comparisons

2. 3. 4.
1.
Query the Include new Adapt
Identify goals
MCM repository elements & Adopt

Source: Buglione L., Gresse von Wangenheim C., Hauck J.C.R., Mc Caffery F., The
LEGO Maturity & Capability Model Approach, Proceedings of 5WCSQ, 5th World
Congress on Software Quality, Shanghai (China), Oct 31- Nov 4 2011

17

Experiencing LEGO... Applying LEGO to RE

Here the four LEGO steps and related activities and outcomes:
1. Identify Goals
 Improve the estimation capability and results by a refinement in the overall management of
requirements (business, technical)
 Assumed the target BPM (Business Process Model) to improve is generically the CMMI RD
process area
1. Query the MCM repository
 Considered RE as the summation of CMMI RM (Req. Management) and RE (Req. Elicitation)
process areas
 Filtered the list of available RE MCMs from the MCM repository
 Next table (EoI – Element of Interest) is a filter of the elements by each of the RE MCMs
considered

1. Include new elements into the target BPM
 Next table (Suggested Improvements) lists the possible EoI matched with the CMMI RD
process (both SPs and GPs)

1. Adapt & Adopt
 Map each practice of the improved process to the related internal QMS process(es)
 Validate the mapping results before using it in the daily activities

18

Experiencing LEGO... Step 2 - EoI: Elements of Interest (1/2)

Model/ Framework Elements of Interest (EoI)
IAG RMM • Technology: the introduction of workflow environments for easily sharing
information for keeping requirements could be useful  CMMI-DEV RD GP2.3
(Elaboration section in Part 1)
• Staff competency: suggested the introduction of Bloom’s levels as
informative notes for all GP 2.5, not only for those two PAs
PRTM CRMM • Level 1: link between product and customer requirements, using e.g. QFD
(quality function deployment)  it could be introduced also in CMMI-DEV RD SP
3.4, not only in SP 2.1 (as currently done) for closing the analysis
BTH REPM • RE.SI (Stakeholders and Req. Source Identification)  more specific
practice to be added about Requirement Elicitation to CMMI-DEV RD SG1
• RE.GA.a2 (Qualify and Quantify Quality Requirements)  currently
missing a more clear and direct link with CMMI-DEV PP SP 1.2
• DS.GA.a2 (Define Requirement Attributes)  currently less stressed (e.g.
FUR vs NFR for FSM/FPA – Function Point Analysis, as requested in CMMI-DEV
PP, SP 1.4

R-CMMi • ML2: P20: Institute Process to Maintain Stability within Project 
always about the need to minimize ‘volatility’, in terms of management  same
comment than for REAIMS practice 9.8

19

Experiencing LEGO... Step 3 - EoI: Elements of Interest (2/2)

Model/ Framework Elements of Interest (EoI)
REAIMS Process MM Basic practices:
• 3.1 Define a standard document structure: missing, could be added in
CMMI-DEV RD SG1, stressing the need for having an organizational ‘standard’ for
comparing different types of requirements, having impact also on planning
(different roles, productivities and schedules for different activities  PP SP 1.4).
Again, it’d help also PP SP 1.2 because it’d address better the
• 3.8 Make the document easy to change  criteria for writing better
requirements, could be stressed more in CMMI-DEV RD SG1 / RM SG1, SP 1.3
• 6.2 Use language simply and concisely  criteria for writing better
requirements could be added as a note for CMMI-DEV RD SP 1.2, sub-practice
#1

Advanced practices:
• 9.8 Identify volatile requirements: suggested to introduce the concept of
‘volatility’ also in the RD process definition by an informative note (e.g. “…
verifying the new need will not be yet addressed by a formalized requirement…”,
with a link to RM, SP 1.3),  see also R-CMMi P20 process, same issue

R-CMM • ML2: P19: Agree and document technical and organisational
attributes specific to project  CMMI-DEV RD deals with customer and
product requirements, not addressing with further informative notes about which
could be possible ‘constraints’ such as those ones from the analysis of
organizational attributes  reinforce RD SP 1.1

20

Experiencing LEGO... Step 3 - Suggestions for Improvement (1/4)

CMMI-DEV v1.3 RD process Suggested Improvements
SG 1 Develop Customer Needs • Introduce a new SP 1.0 about Stakeholders Identification and
Engagement. Rationale: reinforce current formulation, before
running SP 1.1. Nowadays, stakeholder engagement is the sub-
practice #1 within SP 1.1.
• Insert a note about possible standards (de jure/de facto) that
could be consulted/useful for a better application of RD process
(e.g.. (AccountAbility, 2011)).
SP 1.1 Elicit Needs • Introduce a sub-practice about the definition of requirement
attributes, inserting a cross-link with PP SP 1.2 for the
classification of work products (by attribute) to be sized.
• Modify the current WP into: ‘results of requirement elicitation
activities by entity and attribute’ (see previous comment)
SP 1.2 Transform Stakeholders • Rephrase and make more general sub-practice #2: not only
needs functional vs. quality (non-functional) attributes, but possibly
establish all valuable, possible requirements taxonomies and
classifications for the organization (by other criteria)

21


SG 2 Develop Product • Introduce a note within the SG text about the need and
Requirements relevance of define a (standard) document structure (in terms
of ‘documentability’) and suggest – as informative note – some
possible criteria to follow and appraise (e.g. readability, simple
and concise language for writing requirements, etc.).
SP 2.1 Establish Product and • Sub-practice #3: refine the Example box, do no mention generic
Product components quality attributes, but be more specific about requirement
classifications (e..g ISO/IEC 14143-1:1998  functional,
quality, technical)  cross-link with PP 1.2 about attributes for
sizing.
SP 2.2 Allocate Product Components • ---
SP 2.3 Identify Interface • ---
Requirements

22


SG 3 Analyze and Validate • ---
Requirements
SP 3.1 Establish Operational • ---
Concepts and Scenarios
SP 3.2 Establish a Definition of • ---
Required …
SP 3.3 Analyze Requirements • ---
SP 3.4 Analyze Requirements to • Introduce an informative note about the possible usage of QFD
Achieve Balance matrices also here, not only for eliciting and determining
requirements in SP 2.1
SP 3.5 Validate Requirements • ---

23


GP 2.3 Provide Resources • General: stress the need and opportunity from workflow
environments for an easier sharing of information among
stakeholders, whatever the (CMMI) process
• Specific (RD Elaboration): specific need because RD is the
starting process for gathering needs to be translated into
solutions
GP 2.5 Train People • General: introduce the application of the six Bloom’s cognitive
levels (Bloom et al., 1956) for classifying knowledge (see also
IEEE SWEBOK – www.computer.org/swebok)
• Specific (RD Elaboration): add ‘stakeholder engagement’
(AccountAbility, 2011) and ‘requirement sizing’ (ISO, 2011)
GP 2.8 Monitor and Control the • Specific (RD Elaboration): introduce at least one measure about
Process the effectiveness of RD SG1 goal (e.g. % of proposed vs
validated requirements)

24


• Introduction
– Being…Glocal!
Models (MCM)
– The MM-Mania
• Q&A

25

The LEGO Strategy Conclusions & Next Steps

• Requirement Engineering (RE) & Estimation Process
 The main input for an estimate is the list of things to do/accomplish with (requirements)
 The quality (granularity/detail) and quantity of requirements can make the difference
 Typical trend to underestimate and do not be on time/budget/quality
• RE & Maturity/Capability Models (MCMs)
 MCMs can represent a library of best practices for keeping advantange
 Many ‘vertical’ RE MCMs  filtered 7 models (6 staged, 1 continuous) + CMMI RD as the
target process
 The LEGO (Living EnGineering prOcess) approach
 http://slideshare.re/nssLR8 [5WCSQ, Shangai, Nov 2011]
 Choose and integrate the ‘pieces of the puzzle’ you need for your goals  the target is your
QMS, not the model(s) you’re using
 Applying LEGO to RE
 A process (RD) from an horizontal MCM (CMMI-DEV) can be improved with ‘bricks’ from
vertical RE MCM
 Several refinements about requirements scope and granularity, helping in reducing estimation
errors from the initial identification of the amount of activities to work on and improving the
communication among stakeholders  towards a SCRUM-like partnership between client and
provider
 ...try and see!

The difference between the almost right word and
the right word is really a large matter (Mark Twain)
26


• Introduction
– Being…Glocal!
Models (MCM)
– The MM-Mania
• Q&A

27

Hybriding CMMI... Lessons Learned...

www.dilbert.com
URL:
28

Hybriding CMMI... Q&A

Grazie per la vostra attenzione!
Thanks for your attention!
29

Hybriding CMMI... Our Contact Data

Luigi Jean Carlo R. C. Gresse von Fergal
Buglione Hauck Wangenheim McCaffery
Engineering.IT/ETS UFSC UFSC DKIT
luigi.buglione@eng.it jeanhauck@gmail.com gresse@gmail.com fergal.mccaffery@dkit.ie

30

Hybriding CMMI and Requirement Engineering Maturity & Capability Models - Applying the LEGO Approach for Improving Estimates

Recommended

Recommended

More Related Content

What's hot

What's hot (16)

Viewers also liked

Viewers also liked (20)

Similar to Hybriding CMMI and Requirement Engineering Maturity & Capability Models - Applying the LEGO Approach for Improving Estimates

Similar to Hybriding CMMI and Requirement Engineering Maturity & Capability Models - Applying the LEGO Approach for Improving Estimates (20)

More from Luigi Buglione

More from Luigi Buglione (20)

Hybriding CMMI and Requirement Engineering Maturity & Capability Models - Applying the LEGO Approach for Improving Estimates