Ubiquitous Process: An Opportunity or...

Ubiquitous Process: An Opportunity or

Temptation

Ross Jeffery

1

NICTA Copyright 2011 From imagination to impact

Goals

•  To trace the opportunities and context dependent pitfalls for software process research by an investigation of: 1. Knowledge development 2. Process Research examples 3. Research communities 4. Software Process Publications

2


Contents 1. Introduction

2. Models of scientific change

3. Example 1 – Theory Development

4. The current software picture

5. Research Examples

6. Pasteur’s Quadrant and Process Research

7. Communities and Networks

3

8. Conclusions


2. Scientific Change

A number of refinements since the 1970’s but the message remains and the challenge is to understand and make use of the interplay between revolution and evolution in knowledge discovery.

4


Scientific Change

•  1970 – Kuhn proposed that incremental progress is disrupted by upheaval and shift between alternate “paradigms”.

•  1975 – Mulkay et al. proposed a gradual model of change in which upheaval is a special case.

•  “The problem of choice” – Ziman, 1987 Strong tendency for researchers to persist in the same problem area. New research communities may arise from creative tensions.

•  1982-1990 – “Constructivist” studies emphasize social and cultural aspects

5


3. Example 1 - The Early Days in Cost Modeling

•  1970’s Programming productivity Nelson, Johnson, Chrysler, Lawrence

•  Modeling – productivity/size relationship •  Issues: Lack of consistent definition of variables Absence of social and organizational factors Absence of program factors like complexity, quality, documentation completeness, etc.

6


“Peopleware” – DeMarco & Lister (1987)

Effort estimate prepared by:

Average productivity Number of projects

Programmer alone 8.0 19

Supervisor alone 6.6 23

Programmer and supervisor

7.8 16

Data from Jeffery and Lawrence, Journal of Systems and Software, 1985.

7








7.8 16

Systems Analyst 9.5 21

Data from Jeffery and Lawrence, Journal of Systems and Software, 1985. 8








7.8 16

Systems Analyst 9.5 21

(No estimate) 12.0 24

Data from Jeffery and Lawrence, Journal of Systems and Software, 1985. 9


& then: Boehm -Software Engineering Economics

•  Variable definition •  Program factors •  Organizational factors •  Integrated software support environment

10


Schedule Compression in the process

•  COCOMO 1981 •  RCA 1978 •  Putnam 1978

and others……

•  Some model inconsistencies, uncertain data support, Brooks law robustness.

11


Case 1. Avoiding Brooks law

SCATS is a computer system that monitors and controls traffic signals in real-time.

SCATS is an acronym for Sydney Co-ordinated Adaptive Traffic System. It was first developed by the New South Wales Roads and Traffic Authority (RTA). The system is used in more than 80 cities around the world.

12


Empirical Observations

•  The background – Brooks Law (adding more people to a late project will make it even later)

•  Empirical observation - The RTA was an exception to Brooks law (and approx. 20% higher productivity than any other organization of the >30 studied)– why?

•  Empirical observation - The study of 47 software projects showed elapsed time compression and extension with both increased and reduced productivity. Another exception to Brooks Law - why?

13


Case 2 - Data from 47 industry projects concerning schedule and effort (Jeffery - IEEE TSE)

x

x

x

x x

x x x

x x x

x

x x x

x x

x x x x x x

x x x

x

x

x

x

x x

x

x

x x x

x

x

x

Elapsed time ratio Actual/expected

Effort ratio Actual/expected

1.0

1.0 2.1

4.2

Elapsed time compression Elapsed time expansion

14


Compression models and data

x

!"

!"

!"!"

!"!" !"

!"!"!"

!"

!"!"!"

!"!"

!"!"!"!"!"!"

!"!"!"

!"

!"

!"

!"

!"!"

!"

!"

!" !" !"

!"

!"

!"

Elapsed time ratio Actual/expected

Effort ratio Actual/expected

1.0

1.0 2.1

4.2

#$#$%$"

&#'" ()*+,-"

15


Theory development - Staffing model, size and productivity relationships (Jason He Zhang)

SP: Project size RCO: Communication overhead rate SC: Completed size REXD: Experienced employee development rate SP: Remaining size RNWD: New employee development rate RSD: Software development rate RND: Nominal development rate RAS: Assimilation rate PDEX: Experienced employee’s productivity WFT: Total Workforce WFNW: New hired workforce PDNW: New hired employee’s productivity WFET: Experienced workforce for training WFEX: Experienced workforce WFED: Experienced workforce for development

Qualitative abstract structure of staffing process 16


Results

•  A theoretical model using modelling, behavior trees, and simulation show that the empirically observed behavior is possible.

•  The model can then be used to plan and predict the effects on schedule and effort of project staffing patterns

•  In practice - “Information hiding” in the form of staff quarantine was used at the RTA. The key variable is staff numbers and how they can be deployed. Prerequisite – architectural design that allows staff quarantine.

•  A case of observations, laws and theory.

17


Observations, laws and theories

•  Observations – facts or impressions, they tell us “what”. •  Laws – repeatable observations, they tell us “how”. •  Theories – Explain and order or observations, they tell us

“why”. (Endres and Rombach, 2003)

A theory of effort and schedule for software development activity.

18


4. Returning to the bigger picture

Slower development cycles Waterfall methodologies

Focus on cycle time RAD, Spiral.....

Agility in development management – but lots has changed

19


An Industrial observation

•  An American corporation employing over 140,000 people in total.

•  “We're doing some cool things -- advanced agile/lean methodologies on the process side and we're starting to do a lot of mobile development on the product side. We're using all the Atlassian tools and we do a lot of test automation (unit, behavioral, load test) to the point where we're trying to get to a "continuous delivery" process ... the notion that within 8 hours after check in, a piece of code would have been fully vetted automatically in a QA environment and ready for production deployment.”

Email communication July 2010 20


The state of practice •  Management

Agile – penetration of SCRUM etc. as a flexible project management method. Technology support for development. Supported peer review of code and other documents. Test support and automation. Tools support for continuous integration. New deployment technologies

•  Reuse Component assets, open source, and architecture supported services.

•  Process Incremental commitment Systems of systems Formal proofs Business driven

21


State of the art examples

•  Component development •  Services orientation •  Automated testing and continuous integration •  Agile development teams •  Metrics collection and reporting •  Tools and technologies rapidly changing •  Adaptive systems •  Systems of Systems •  Cloud •  Social Networks •  Internet of things

22


Recent Process Research Examples

•  Smart Grid * (Architecture assessment process). Defence - ULS Context

•  LIXI – Loan industry XML initiative (Business Process Context)

•  L4 Verified project

* Len Bass and IEEE Computer March 2011

23


Smart Grid - What is Demand Response?

• The Smart Grid is the electric power grid enhanced with IT to make it more reliable, efficient, and secure.

• Demand Response (DR) is a collection of programs by utility companies whose main objective is to reduce peak load during periods of high demand for electricity or when grid reliability is at risk.

Used with permission - Len Bass “Architecture Evaluation without an Architecture: Experience with the Smart Grid”

24


ULS System Perspective

• Residential DR exhibits aspects of Ultra-Large-Scale (ULS) systems: •  Continuous evolution and deployment. •  Heterogeneous, inconsistent, and changing elements. •  Erosion of the people/system boundary.

25


Some Implications

• Conducted an analysis of architectures for residential DR and found risks, risk themes, and tradeoffs.

• Created a method for analyzing architectures that have yet to be built, or even designed

•  it explores the space of possible architectural options •  it finds potential risks with respect to the achievement of important

system quality attributes (QAs): reliability, performance, usability, interoperability, modifiability, etc.

•  it finds potential design tradeoffs

• This method can be used to help organizations make informed architectural decisions when designing complex under-determined systems.

26


Defence Concerns in ULS space

The response: Defence “ Systems of Systems Integration – Capability Scoping Study” 2011

27


BP: Industry Contributions and Impact

NICTA Contributions to LIXI, 2005 - 2008 •  Reference business process models and handbooks •  LIXI Property Valuations & “Visible Loans” Reference Architecture Adoptions •  Released to 150+ financial institutions

–  Adoption example: property valuation process and design • Adopted by a small Australia valuation company; Scaled from a dozen

in-house valuers to 100+ contract valuers using mobile apps and Web services

–  Adoption example: smart feed technologies and Excel integration • Adopted by software company Solution4: Created more efficient

lending processes without updating legacy systems or investing in middleware

•  Our approach is being extended to support the new National Electronic Conveyancing System and Superannuation 28


BP: Industry Contributions and Impact

•  Before: applying for a mortgage was amazingly complex, involving up to nine separate organisations communicating in various proprietary ways

•  Now: Partial LIXI implementation in nearly all major banks and aggregators (80% loan origination and 50% of the mortgage insurance applications are using LIXI)

•  Before: $1 billion spent on the loan approval process every year in Australia, and almost $200 million of it in re-work.

•  Now: streamlined processes. Saving at least $68 out of $450 (46 million to 107 million savings a year)

•  Example: CBA reduces the time of approving a home loan from 14-22 days to 14-15 mins.

29


BP: Research Contribution and Impact

•  13 publications and continuing •  Venues

–  mostly business process management and software architecture community and book chapters

–  One paper in software process community [4] •  Citation numbers (3 years only): 15 [1], 9[4], 6 [3]

[1] L. Zhu, L. Osterweil, M. Staples, U. Kannengiesser, B. Simidchieva, “Desiderata for Languages to be Used in the Definition of Reference Business Processes” International Journal of Software and Informatics (IJSI) , Volume 1, Issue 1, pp. 37-66, 2008 [2] Kannengiesser, U. and L. Zhu, Concise Architectures for Flexible Business Processes, in Business Process Modeling: Software Engineering, Analysis and Applications. 2010. (book chapter) [3] Zhu, L., M. Staples and V. Tosic, “On Creating Industry-Wide Reference Architectures,” in The 12th IEEE International EDOC Conference (EDOC’08), Munich, Germany, 2008. [4]. L. Zhu, M. Staples, and R. Jeffery, “Scaling Up Software Architecture Evaluation Processes,” in International Conference on Software Process (ICSP’08)

30


Example Publishing system process (XML)

•  Glossary •  Wireframes •  Generation and coding

•  “The customer must understand the technical details. So the language must be understandable to the non-technical.”

31


The process research determinants

•  Software type •  Process type •  Context Goals People Tools Distribution

•  Current Empirical Understanding

32


Pasteur’s Quadrant

http://openeducationresearch.org/2009/01/pasteurs-and-edisons-quadrants/ 33


Basic research

•  Goal – to advance knowledge •  May be a Kuhnian event •  Short term – possibly few citations (depending

on community and network) •  Long term – many citations

34


Applied Research

•  Goal – Technology use •  Within software or anywhere

35


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`'.,"#.#"Z$5*$%$.*E(A5&<.,P%"%5#P@$%5.<G5%)B#H.*$5#)),% 0II1 K02N3 VV59)PE$ O.<.?"<?5#($5G$Z$,)@7$<#5)B5,.*?$5%)B#H.*$5%P%#$7% 0IM1 S0M23 O:5:.,%#$.G 4,$7$<#%5)B5%)B#H.*$5%E"$<E$ 0IMM K0M3/ ]5O$GZ"G)Z"E=&5E,.%%"B"E.#")<5.<G5E)7@.*"%)<5B*.7$H)*Y5B)*5%)B#H.*$5.*E("#$E#'*$5G$%E*"@#")<5,.<?'.?$% /111 S0M/0 9+5["$,G"<? &*E("#$E#'*.,5%#P,$%5.<G5#($5G$%"?<5)B5<$#H)*YFC.%$G5%)B#H.*$5.*E("#$E#'*$% /111 +($%"%

Citation counts for “Software”

36

NICTA Copyright 2011 From imagination to impact 37


Citations by discipline: Average per article

38

From:: Carlo Ghezzi - Reflections on 40+years of software engineering research observed through ICSE: an insider's view, ICSE, 2009 Average citations vary widely by field. (Data from Thomson Scientific)


2005

Year Venue Papers Google Avg. Google Max.

2005 SPW 41 6.7 32

2005 PROFES 42 4.5 30

2005 EuroSPI 18 5.4 25

conference sum 101 5.5 32

journal SPIP 27 25 227

2008

Year Venue Papers Google Avg. Google Max.

2008 ICSP 35 8.2 28

2008 PROFES 31 9.1 21

2008 EuroSPI 18 3.6 23

conference sum 84 7.5 28

journal SPIP 40 4.6 21

39


SPW2005 ICSP2008 PROFES2005 PROFES2008 EuroSPI2005 EuroSPI2008 SPIP2005 SPIP2008 1 3 24 1 11 3 1 227 2 2 18 6 1 8 11 2 4 4 3 7 5 0 7 25 1 13 8 4 2 2 1 6 10 3 7 2 5 3 14 14 3 1 1 6 2 6 15 3 1 14 1 4 5 1 7 21 12 0 7 0 4 5 13 8 3 0 3 14 2 0 196 3

9 31 8 0 5 3 2 12 4 10 0 13 1 1 0 0 30 6 11 1 11 6 11 6 1 14 4 12 10 12 0 10 10 23 4 2 13 5 1 5 7 5 9 36 3 14 2 6 1 18 13 3 10 0 15 0 11 9 13 0 1 20 9 16 1 5 3 14 5 6 1 13 17 2 11 1 13 2 3 8 21 18 1 9 7 21 0 1 5 4 19 1 14 8 9 5 5 20 1 4 0 12 18 0 21 5 2 7 4 13 5 22 2 1 5 1 15 2 23 0 5 3 7 10 3 24 0 5 11 6 4 4 25 10 4 8 10 1 1 26 2 10 1 11 2 6 27 32 5 1 9 5 0 28 8 13 11 10 3 29 23 5 2 7 1 30 7 28 3 5 3 31 1 3 8 7 0 32 0 10 6 0 33 14 10 0 3 34 4 6 5 2 35 1 9 9 1 36 13 5 3 37 11 3 13 38 2 4 9 39 3 30 11 40 2 0 9 41 7 0 42 5

Avg. 6.7 8.2 4.5 9.1 5.4 3.6 25.0 4.6

Conference Overall

2005 5.5 9.7

2008 7.5 6.6

40


Citation

SPIP2008

SPIP2005

EuroSPI2008

EuroSPI2005

PROFES2008

PROFES2005

ICSP2008

SPW2005

40

30

20

10

0

Citation

SPIP2008

SPIP2005

EuroSPI2008

EuroSPI2005

PROFES2008

PROFES2005

ICSP2008

SPW2005

200

150

100

50

0


2005 Conferences

Rank Venue Title Author(s)

Google

(>20)

1 SPW Software Process Management: Practices in China Qing Wang, Mingshu Li 32 2 SPW Integrated Software Process and Product Lines* Dieter Rombach 31

3 PROF

ES Measuring Similarity of Large Software Systems Based on Source Code Correspondence

Tetsuo Yamamoto, Makoto Matsushita, Toshihiro Kamiya, Katsuro Inoue 30

4 EuroS

PI Pair Programming vs. Side-by-Side Programming Jerzy R. Nawrocki, Michal Jasinski, Lukasz Olek, Barbara Lange 25

5 SPW Process Programming to Support Medical Safety: A Case Study on Blood Transfusion

Lori A. Clarke, Yao Chen, George S. Avrunin, Bin Chen, Rachel Cobleigh, Kim Frederick, Elizabeth A. Henneman, Leon J. Osterweil 23

6 SPW Unifying Microprocess and Macroprocess Research* Leon J. Osterweil 21 ( * keynote)

2005 Journal - SPIP Rank Issue Title Author(s)

Google

1 1 Formalizing cardinality-based feature models and their specialization

Krzysztof Czarnecki, Simon Helsen, Ulrich Eisenecker 227

2 2 Staged configuration through specialization and multilevel configuration of feature models


3 3 Process modeling across the web information infrastructure Chris Jensen, Walt Scacchi 36

4 2 An evaluation of CMMI process areas for small- to medium-sized software development organisations F. G. Wilkie, D. McFall, F. McCaffery 30

42


2008 Conferences Rank Venue Title Author(s)

Google

(>15)

1 ICSP Reflections on 10 Years of Software Process Simulation Modeling: A Systematic Review He Zhang, Barbara Kitchenham, Dietmar Pfahl 28

2 ICSP Benefits of Global Software Development: The Known and Unknown*

Par J. Agerfalk, Brian Fitzgerald, Helena Holmstrom Olsson, Eoin O Conchuir 24

3 EuroS

PI A Software Engineering Lifecycle Standard for Very Small Enterprises

Claude Y. Laporte, Simon Alexandre, Rory V. O'Connor 23

4 PROF

ES A Fault Prediction Model with Limited Fault Data to Improve Test Process Cagatay Catal, Banu Diri 21

5 PROF

ES Comparing Assessment Methodologies for Free/Open Source Software: OpenBRR and QSOS Jean-Christophe Deprez, Simon Alexandre 18 ( * keynote)

2008 Journal - SPIP Rank Issue Title Author(s)

Google

(>15)

1 3 Understanding a lack of trust in Global Software Teams: a multiple-case study Nils Brede Moe, Darja Šmite 21

43


Further Detail 2005

44

2005 Conferences

Rank Venue Title Author(s) Googl

e (>20) Self-cited

SP-related

SE-related Others

1 SPW Software Process Management: Practices in China Qing Wang, Mingshu Li 32 23 5 3 1

2 SPW Integrated Software Process and Product Lines* Dieter Rombach 31 0 30 1 0

3 PROFES

Measuring Similarity of Large Software Systems Based on Source Code Correspondence

Tetsuo Yamamoto, Makoto Matsushita, Toshihiro Kamiya, Katsuro Inoue 30 5 1 17 7

4 EuroSPI Pair Programming vs. Side-by-Side Programming

Jerzy R. Nawrocki, Michal Jasinski, Lukasz Olek, Barbara Lange 25 3 18 3 1

5 SPW Process Programming to Support Medical Safety: A Case Study on Blood Transfusion

Lori A. Clarke, Yao Chen, George S. Avrunin, Bin Chen, Rachel Cobleigh, Kim Frederick, Elizabeth A. Henneman, Leon J. Osterweil 23 11 1 8 3

6 SPW Unifying Microprocess and Macroprocess Research* Leon J. Osterweil 21 3 16 0 2 ( * keynote)

2005 Journal - SPIP

Rank Issue Title Author(s) Google

(>20)

1 1 Formalizing cardinality-based feature models and their specialization


2 2 Staged configuration through specialization and multilevel configuration of feature models


3 3 Process modeling across the web information infrastructure Chris Jensen, Walt Scacchi 36 21 9 1 5

4 2 An evaluation of CMMI process areas for small- to medium-sized software development organisations F. G. Wilkie, D. McFall, F. McCaffery 30 5 24 0 1


Citations

45

The sociology of citations* • a way to carry out a conversation – reference included to show that topic is of interest to someone else or to prove that the author knows the literature – the cited explains some result, not necessarily of the cited author – the cited represents another approach, or is an example of…

* From: Carlo Ghezzi - Reflections on 40+years of software engineering research observed through ICSE: an insider's view, ICSE, 2009


Conversations: Communication of innovation

•  Diffusion is influenced by organization and network characteristics and individual characteristics.

•  Diversity is normally reduced and this will increase citations.

•  Emerging networks will have publications more widely dispersed, less visible and less cited.

46

Perry & Rice, in Science Communication, 21,1, September, 1999.


Developing research networks

•  Networks are based on self interest, personal benefit.

•  Communities of practice are for self interest and for the benefit of others in the community.

•  Build the community - “Community development is a core requirement for a successful research network.”

•  Engage critically - research methods, tools, issues and trends, individual difference.

47

Eyman, Sheffield & DeVoss, in Computers and Composition, 26, 2009.


Community Example: Empirical

•  ISERN 1993, Basili, Torii, Rombach, Cantone, Votta, Jeffery, et al.

•  Metrics 1993 •  EMSE, 1996, •  ISESE/ESEM, 2002

48


Software and Systems Process

•  ISPW, 1984 •  EuroSPI, 1994 •  Profes, 1999 •  Journal of Software Maintenance and

Evolution: Research and Practice incorporating SP:IP (1995)

49


Choice: Process continuum

•  Software and systems development processes

•  Software supported processes •  Nothing to do with software, but still processes

50


Choice: Software and Systems Process Targets Continuum

•  Lifecycle Processes •  Process modeling and analysis, process

refinement, process improvement •  Domain specific lifecycles and newer

technologies

51


Temptations

•  To not be a community. •  To not engage critically. •  To apply existing knowledge to non-expert

research questions. •  To only do applied research. This is likely a

bonus, not core.

52


Conclusions

•  Software and Systems process knowledge arises from basic and use-inspired basic research

•  The cost modeling example showed use-inspired basic research leading to theory through observation and modeling.

•  The current industrial setting (use inspiration) provides many opportunities for use inspired basic research adaptive systems process systems of systems cloud smart grid financial systems publishing, etc. etc…….

53


Conclusions

•  Process research has had impact in that citations are relatively high and industrial practice has been changed.

•  Conversations are dependent on: Community Critical engagement Basic and use-inspired basic research.

•  The variations in software type, process type, and context allow for rich research opportunities.

54


Conclusions

•  The research space in software and systems process is too large for broad scale one size fits all solely academic research inroads. SofS Financial/business Publishing Embedded software Defence etc. etc. etc.

•  The software and systems processes will differ significantly and therefore the research opportunities are large.

•  We must have community.

55


Questions?

56

Date post:	08-Jun-2018
Category:	Documents
Upload:	hoanganh
View:	217 times
Download:	0 times

Ubiquitous Process: An Opportunity or...

Documents