Postdoc Symposium - Abram Hindle

791 views

Published on

Evidence-based Software Process Recovery: A Post-doctoral View

Published in: Technology
0 Comments
1 Like
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total views
791
On SlideShare
0
From Embeds
0
Number of Embeds
9
Actions
Shares
0
Downloads
3
Comments
0
Likes
1
Embeds 0
No embeds

No notes for slide

Postdoc Symposium - Abram Hindle

  1. 1. Evidence-based Software Process Recovery: A Postdoctoral View Evidence-based Software Process Recovery: A Postdoctoral View Abram Hindle Department of Computing Science University of Alberta Edmonton, Alberta Canada http://softwareprocess.es/ abram.hindle@softwareprocess.es 1Abram Hindle
  2. 2. What are we going to do? Theory Practice ?Business ModelingRequirementsAnalysis & DesignImplementationTestDeploymentCM and SCSProject MangementEnvironment 2
  3. 3. 3
  4. 4. Motivation: Stakeholders Fixers or Investors andStar Programmers Acquisitions Managers Employees assigned New Developers to a ISO9000 conformance project 4
  5. 5. Proposed Process Recovered ProcessWorkflows Is my proposed process actually being used? 5
  6. 6. Proposed and Recovered Differences between Process Overlayed Proposed and RecoveredWorkflows I can compare and contrast the observed process versus the expected process! 6
  7. 7. How to get an overview: Interviews time consuming annoying ERDANG access 7
  8. 8. Cant we just Revisions Software summarize Repositorieswhat is going Source Codeon within this Source Code Build / Configuration project? Tests Documentation Phases Disciplines Inception Elaboration Construction Transition Business Modeling Requirements Analysis & Design Implementation Test Deployment CM and SCS Project Mangement Environment Initial Elab Elab Const Const Const Trans 8
  9. 9. Research Relationships Software Intent, Purpose Behaviour Development and Tasks Process Release Patterns: Source/Test/Build/Documentation Topic Analysis Large Changes Recovered Unified Process Views 9
  10. 10. Mining Software Repositories Revisions Source Code Source Code Build / Configuration Tests Documentation 10
  11. 11. Source Acquisition Version Control Revisions discussions bugs sourceInitial Repositories and artifacts 11
  12. 12. SOFTWAREPROCESSRECOVERY 12
  13. 13. Release Patterns: STBD Release Event Linear Regression Source Source Code Revisions per Time Unit (day) Smoothed Summed Near ReleaseRevisionsper Time Test Test RevisionsUnit per Time Unit (day)summed Smoothedper time Summed Near Releaseunitbefore and Build Build Revisionsafter a per Time Unit (day)release Smoothed Summed Near Release Documentation Documentation Revisions per Time Unit (day) Smoothed -n time units 0 +n time units Summed Near Release Time [Hindle ICSM07] 13
  14. 14. STBD applied to SQLiteSQLITE 14
  15. 15. Proportion of Commits 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 C 0.45 or re ct iv e A d ap ti ve P er fe c ti ve S No o n C urc - od e e - m Im en p ta le- ti on 15 Maintenance Classes of Large Changes[Hindle ICPC09]
  16. 16. What is this commit about? Added a test for bug #1326 on OSX 16
  17. 17. What is this commit about? Added a test for bug #1326 on OSX 17
  18. 18. What is this commit about? Added a test for bug #1326 on OSX Maintain-Reliability ability Portability 18
  19. 19. But we have many commits.. Maintain-Reliability Portability ability 19
  20. 20. Cross Project Relevance Version Version Control Control efficiency Shared usability reliability and functionality Concepts (includes correctness) maintainability Version portability Version Control Control 20
  21. 21. Quality RelatedNon functional requirementsportability reliability and functionality (includes correctness) usability efficiencymaintainability [cleland-huang03] [ernst10] 21
  22. 22. Word Bag Examples Portability Reliability portability reliability transferability failure interoperability error documentation redundancyinternationalization fails i18n bug ... ... 22
  23. 23. Developer Topics Commit Commit Developer Topic Developer Topic purpose?Maintainability Reliability LDA 23
  24. 24. Unique Topics Labelled Developer Topics Time (months) 24
  25. 25. Labelled Developer Topics LinuxUnique Topics Kernel Windows AMD64 Time (months) 25
  26. 26. Labelled Developer Topics efficiencyportability efficiency portability functionalit maintainability efficiencyUnique Topics reliability maintainability portability functionality Time (months) 26
  27. 27. MaxDB 7.500 TimelineMaintainability Maintainability MaintainabilityPortability Portability Effeciency Reliability Effeciency 27 [MSR 2011]
  28. 28. SOFTWAREPROCESSRECOVERY: 28
  29. 29. Recovered Unified Process Views Theory PracticeBusiness ModelingRequirementsAnalysis & DesignImplementationTestDeploymentCM and SCSProject MangementEnvironment [ICSM10] 29
  30. 30. UP Requirements Signal + + = 30
  31. 31. UP Implementation Signal = 31
  32. 32. UP Testing Signal = 32
  33. 33. SQLite Case Study 33
  34. 34. SQLite Case Study: 2009 TestingBusiness Modelling DeploymentRequirements Config/SCSAnalysis Project ManagementImplementation Environment 34
  35. 35. UP Observability? ? ? Disciplines Business Modeling Requirements Analysis & Design Project Mangement Environment 35
  36. 36. Future Work ?People Accuracy ? Unknown Project and MSR Benchmarkteams ValidationIndustrial Iteration Identification 36
  37. 37. Common Threads Idioms *.doxygen *.tex FILES Makefile Makefile.* External INSTALL doc/ READMEAUTHORS TODO configure configure.* Project*.txt Documentation setup.hs *test* build.xml setup.py Language *.hs *.ml *.php *.java *.scm shared *.lisp *.cpp *.tcl *.c *.C Source Code *.py *.t unit tests *.sql vocabularySource Code *.pl Source Code *.test *.pl *.pm *.rb Test vocabulary vocabulary Shared Terms Project Project Usability Maintainability Internal Project Portability Reliability Efficiency Language 37
  38. 38. Process Recovery Summary Disciplines Business Modeling Requirements Analysis & Design Implementation Test Deployment CM and SCS Project Mangement ManagersSoftware Process Environment Maintenance Classes Release Patterns DeveloperportaTopic Analysis efficiency bility Fixers or 0.45 and Labelling 0.4 0.35 efficiency Star Programmers Recovery Source Proportion of Commits 0.3 0.25 reliability 0.2 0.15 Test 0.1 maintainability portability 0.05 functionality Build 0 LDA ta le- on e e ve C urc - iv iv n e - en p ti od e LSI S No ti m Im ct ct Documentation ap re fe o d or er [Hindle09ICSM] [hindl A [Hindle ICPC09] C P [Hindle and Ernst http://softwareprocess.es/name/] [Hindle ICSM07] Time New Developers Tests Source Documentation Usability Maintainability Reliability Portability Efficiency Build Investors and Acquisitions Version Control ISO9000 Consultants Revisions Discussions Bugs Source 38
  39. 39. Lessons Learned Boring Slides Conferences!Collaborate!Organization A PhD is not enough 39

×