Agile and Lean Software Development


Published on

My presentation at VTU's Faculty Development Program (FDP) organised by Computer Society of India (CSI) Bangalore on Jan 11, 2013.

Published in: Technology
1 Comment
No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide
  • The five-step thought process for guiding the implementation of lean techniques is easy to remember, but not always easy to achieve:Specify value from the standpoint of the end customer by product family.Identify all the steps in the value stream for each product family, eliminating whenever possible those steps that do not create value.Make the value-creating steps occur in tight sequence so the product will flow smoothly toward the customer.As flow is introduced, let customers pull value from the next upstream activity.As value is specified, value streams are identified, wasted steps are removed, and flow and pull are introduced, begin the process again and continue it until a state of perfection is reached in which perfect value is created with no waste.
  • Agile and Lean Software Development

    1. 1. Agile and LeanSoftware Development Tathagat Varma Sr. Director Yahoo!
    2. 2. Discussion Topics• SDLC• Waterfall• Agile• Lean• Kanban• Lean Startup
    3. 3. the world around us…yesterday!
    4. 4. Microsoft Windows timeline
    5. 5. Other major OS and tools timeline
    6. 6. A typical support timeline
    7. 7. As a contrast, what are consumer internet companies doing?• Continuous Integration -> Continuous Delivery -> Continuous Deployment• On „good days‟, Flickr releases a new version every half an hour (Jun 20, 2005)• IMVU pushes a revision of code to the website every nine minutes (Feb 10, 2009)• The other day we passed product release number 25,000 for WordPress. That means we‟ve averaged about 16 product releases a day, every day for the last four and a half years! (May 19, 2010)• A new version of Google Chrome now due every six weeks (Jul 22, 2010)• Facebook does code push twice a day (Aug 4, 2012)
    8. 8. adoption @ net speed!!!
    9. 9. darwin at work on internet 2006 2009
    10. 10. yet, our product development sucks!
    11. 11. Let’s understand the ‘craft’ first… • Sheer joy of making things • Pleasure of things that are useful to other people • Fascination of fashioning complex puzzle-like objects of interlocking moving parts and watching them work in subtle cycles • Joys of always learning, which sprints from the non- repeating nature of the task • Delight of working in such a tractable mediumThe Mythical Man Month – Fred Brooks, 1975
    12. 12. software development life cycles• Ad-hoc• Serial• Iterative• Incremental• Iterative/Incremental
    13. 13. risk management in SDLCs
    14. 14. notion of ‘progress’ vs. time
    15. 15. Waterfall Model• Wrongly inspired by assembly-line manufacturing processes of the day• Economics supported “measure twice, cut once” leading to up-front planning and BDUF• Single-pass, sequential process with hand-offs and feedback loops between adjoining phases• Transition to next phase only upon completion of current phase
    16. 16. Waterfall Software Development Limitations and Assumptions 1. Wrong analogy: Software development ≠ Production 2. Customers know EVERYTHING upfront and that requirement won’t change 3. Legacy from the past: implicitly assumes CPU time is costly, so focuses on doing everything upfront to minimize ‘machine time’ for trial and error 4. “Wicked Problem”: Designers and developers know how exactly how to build 5. Very long feedback cycles not suitable for today’s pace of innovationPicture from
    17. 17. As a result, software is… Costly Buggy Late
    18. 18. and the costs…?
    19. 19. Holy Grail of Software Development • Better: higher quality, more reliability, higher performance, more usable… • Faster: speedier development • Cheaper: no budget overruns
    20. 20. But the reality?
    21. 21. Preamble to Agile MovementSoftware Crisis, 1965-85: The major cause of thesoftware crisis is that the machines have becomeseveral orders of magnitude more powerful! To putit quite bluntly: as long as there were nomachines, programming was no problem at all;when we had a few weak computers, programmingbecame a mild problem, and now we have giganticcomputers, programming has become an equallygigantic problem. — Edsger Dijkstra, The HumbleProgrammer
    22. 22. Software CrisisThe causes of the software crisis were linked to theoverall complexity of hardware and the softwaredevelopment process. The crisis manifested itself inseveral ways:• Projects running over-budget.• Projects running over-time.• Software was very inefficient.• Software was of low quality.• Software often did not meet requirements.• Projects were unmanageable and code difficult to maintain.• Software was never delivered.
    23. 23. and the response?Frameworks, Standards and Certifications
    24. 24. …and the result?…good start…
    25. 25. …but poor finish!
    26. 26. Why?The World had changed! Software increasinglybecame more business-critical, and delays, costoverruns and poor quality were significantly lessacceptable then before!• Process: Long-lead development process ineffective in a dynamic and global world• Management: Command and control model unsuitable for fostering collaboration required to solve complex problems• Technology: Advancements in computers, compiler technology and debugging and testing tools greatly improved the economics of software development
    27. 27. What is the most important part in these two machines? “The Brakes!!!” They let you go faster…
    28. 28. Agility vs. Discipline?
    29. 29. Advent of Agile and Lean Methodologies• 1970: Royce critiques Waterfall and offers improvement ideas• 1986: Barry Boehm proposes Spiral Model• 1971: Harlan Mills proposes Incremental Development• 1987: Cleanroom Software engineering• 1991: Sashimi Overlapping Waterfall Model• 1992: Crystal family of methodologies• 1994: DSDM• 1995: Scrum• 1996: Rational Unified Process framework• 1997: Feature Driven Development• 1999: Extreme Programming Explained• 2001: Agile Manifesto is born• 2003: Lean Software Development• 2005: PM Declaration of Interdependence• 2007: Kanban-based software engineering• 2008: Lean Startup• 2009: Scrumban• 20xx: Something new !?! (hopefully!)
    30. 30. What is agile really all about? • Empowered individuals • Collaboration Motivated • Democratic decision-making Individuals and transparency Self-organizing • Shorter feedback cycle x- • Manage changing prioritiesfunctional • Increased productivity Teams Agile Businesses • Higher ROI • Faster time to market • Better User Experience
    31. 31. Why is it so hard?52 39 34 %Organizational Culture % Resistance to Change % Management Support
    32. 32. test-code-refactor loop
    33. 33. feedback loop in agile lifecycles
    34. 34. agile lifecycle – big picture
    35. 35. from daily builds to project
    36. 36. Scrum
    37. 37. What’s happening here?
    38. 38. Feedback Loops in Traditional Techniques vs. Agile Techniques
    39. 39. Agile Development Value Proposition
    40. 40. Does Agile work?
    41. 41. What is Lean?• The core idea is to maximize customer value while minimizing waste. Simply, lean means creating more value for customers with fewer resources.• A lean organization understands customer value and focuses its key processes to continuously increase it. The ultimate goal is to provide perfect value to the customer through a perfect value creation process that has zero waste.
    42. 42. Lean Thinking• Lean thinking changes the focus of management from optimizing separate technologies, assets, and vertical departments to optimizing the flow of products and services through entire value streams that flow horizontally across technologies, assets, and departments to customers.• Eliminating waste along entire value streams, instead of at isolated points, creates processes that need less human effort, less space, less capital, and less time to make products and services at far less costs and with much fewer defects, compared with traditional business systems. Companies are able to respond to changing customer desires with high variety, high quality, low cost, and with very fast throughput times. Also, information management becomes much simpler and more accurate.
    43. 43.
    44. 44. Lean Principles 1. Identify Value 2. Map 5. Seek the ValuePerfection Stream 4. 3. Create Establish Flow Pull
    45. 45. Wastes in Lean
    46. 46. How they manifest?
    47. 47. Lean Software Development • Optimize the Whole • Eliminate Waste • Build Quality In • Learn Constantly • Deliver Fast • Engage Everyone • Keep Getting
    48. 48. Optimize the WholeOptimizing a part of a system will always, over time,sub-optimize the overall system.Focus on the Entire Value StreamFrom concept to cash.From customer request to deployed software.Deliver a Complete ProductCustomers dont want software; they want their problemssolved. Complete solutions are built by complete teams.Think Long TermBeware of governance and incentive systems that drive shortterm thinking and optimize local performance.
    49. 49. Eliminate WasteWaste is anything that does not add customer value.The three biggest wastes in software development are:Building the Wrong Thing"There is nothing so useless as doing efficiently that which should notbe done at all." –Peter DruckerFailure to LearnMany of our policies – for example: governance by variance fromplan, frequent handovers, and separating decision-making from work –interfere with the learning that is the essence of development.ThrashingPractices that interfere with the smooth flow of value –taskswitching, long lists of requests, big piles of partly done work – deliverhalf the value for twice the effort.
    50. 50. Wastes in Software DevelopmentWastes in Manufacturing Wastes in Software DevelopmentInventory Partially done workExtra Processing Extra processesOverproduction Extra featuresTransportation Task SwitchingWaiting WaitingMotion MotionDefects Defects
    51. 51. Build Quality InIf you routinely find defects in your verification process, your process isdefective.Final Verification Should Not Find Defects!Every software development process ever invented had as its primarypurpose to find and fix defects as early in the development process aspossible.Mistake-Proof your Process with Test-First DevelopmentTests – including, unit tests, end-to-end tests, and integration tests –must be available to establish confidence in the correctness of thesystem at any timeduring development, at every level of the system.Break DependenciesSystem architecture should support the addition of any feature at anytime.
    52. 52. Learn ConstantlyPlanning is useful. Learning is essential.Predictable Performance is Driven by FeedbackA predictable organization does not guess about the future and call it aplan; it develops the capacity to rapidly respond to the future as itunfolds.Maintain OptionsThink of code as an experiment – make it change-tolerant.Last Responsible MomentLearn as much as possible before making irreversible decisions. Dontmake decisions that will be expensive to change before their time –and dont make them after their time!
    53. 53. Learn FastStart with a deep understanding of all stakeholders and what they willvalue. Create a steady, even flow of work, pulled from this deepunderstanding of value.Rapid Delivery, High Quality, and Low Cost are Fully CompatibleCompanies that compete on the basis of speed have a big costadvantage, deliver superior quality, and are more attuned to theircustomers needs.Queuing Theory Applies to Development, not Just ServersFocusing on utilization creates a traffic jam that actually reducesutilization. Drive down cycle time with small batches and fewer things-in-process. Aggressively limit the size of lists and queuesManaging Workflow is a lot easier than Managing SchedulesThe best way to establish reliable, predictable deliveries is to establishreliable, repeatable workflows with iterations or a kanban system.
    54. 54. Keep Getting BetterResults are not the point – the point is to develop the people and thesystems capable of delivering results.Failure is a Learning OpportunityThe most reliable performance comes when even small failures aredeeply investigated and corrected; when noise is not tolerated.Standards Exist to be Challenged and ImprovedEmbody the current best known practice in standards that everyonefollows, while actively encouraging everyone to challenge and changethe standards.Use the Scientific MethodTeach teams to: establish hypotheses, conduct many rapidexperiments, create concise documentation, and implement the bestalternative.
    55. 55. Engage EveryoneThe time and energy of bright, creative people are the scarce resources intodays economy, and the basis of competitive advantage.People who are paid fairly and adequately are motivated byautonomy, mastery, and purpose.AutonomyThe most effective work groups are semi-autonomous teams with an internalleader with end-to-end responsibility for complete, meaningful tasks.MasteryRespect for people means providing the challenge, feedback, andenvironment that enables everyone to become excellent.PurposeTie work to value. Only by believing in the purpose of their work will peoplebecome engaged in achieving that purpose.
    56. 56. Kanban System
    57. 57. How Kanban helps achieve “Just- in-Time” • For example, to efficiently produce a large number of automobiles, which can consist of around 30,000 parts, it is necessary to create a detailed production plan that includes parts procurement. Supplying "what is needed, when it is needed, and in the amount needed" according to this production plan can eliminate waste, inconsistencies, and unreasonable requirements, resulting in improved productivity.
    58. 58. Kanban for Software • Visualize the Workflow: Represent the work items and the workflow on a card wall or electronic board • Limit Work-in-Progress (WIP): Set agreed upon limits on how many work items are in progress at a time • Measure and Manage Flow: Track work items to see if they are proceeding at a steady, even pace • Make Process Policies Explicit: Agree upon and post policies about how work will be handled • Use Models to Evaluate Improvement Opportunities: Adapt the process using ideas from Systems Thinking, Deming, etc.Kanban: Successful Evolutionary Change for your Technology Business – David Anderson
    59. 59. Why Kanban in Software Engineering? Don’t build Don’t write Don’t write Don’t test features that more specs more code more code nobody than you can than you can than you can needs right code test deploy now
    60. 60. Lean Startup PrinciplesThe Lean Startup method teaches you how todrive a startup-how to steer, when to turn, andwhen to persevere-and grow a business withmaximum acceleration.• Entrepreneurs are everywhere• Entrepreneurship is management• Validated Learning• Innovation Accounting• Build-Measure-Learn
    61. 61. Build- Measure-Learn Loop
    62. 62. does iterating help?
    63. 63. are small teams more productive?
    64. 64. does colocation impact team performance?
    65. 65. is small batch size faster?
    66. 66. Recap• Agile ≠ Faster, but Sooner• Agile ≠ No planning, but Adaptive Planning• Agile ≠ More work, but ‘Done’• Agile ≠ No documentation, but Just Enough• Lean ≠ Cut the corners, but Reduce Waste• Agile or Lean don’t just change the development process, but enable a radical change in organizational culture, leadership and management practices that is more in line with business needs and social values and norms of today
    67. 67. References•••• time.html•• overview/
    68. 68. It’s not about the method!• A photographer went to a socialite party in New York. As he entered the front door, the host said „I love your pictures – they’re wonderful; you must have a fantastic camera.‟• He said nothing until dinner was finished, then: „That was a wonderful dinner; you must have a terrific stove.‟ – Sam Haskins
    69. 69. Connect Blog: Email: Slides: Twitter: Articles: