What lies beneath
â˘
6 likesâ˘1,741 views
1) EventStorming is a technique that uses modeling with sticky notes, markers and paper to surface domain events and capture hotspots in a system. It enables cross-perspective conversations and can lead to unexpected insights. 2) Analyzing bottlenecks is important because most bottlenecks are found in business processes, policies, user interfaces and read models. However, simply documenting a bottleneck may not solve the problem, as organizational culture and emotions also influence whether issues are actually addressed. 3) When modeling legacy systems, it is important to identify bounded contexts and focus modeling and improvements within safer contexts first to minimize risk. Addressing technical issues alone may not be enough, as unaddressed cultural and emotional aspects can hinder
What lies beneath
- 2. About me Very hard to explain my job to my mother running www.avanscoperta.it Modelling (almost) everything with sticky notes, markers and a paper roll. Calling this stuff
- 3. Quick recap of Focusing on Big Picture only
- 4. Big Picture Workshop Invite the right people Provide an unlimited modelling space Surface, Markers, stickies Start Modelling with Domain Events
- 6. Establish a timeline Some facilitator tricks will kickstart the discussion quickly
- 8. Capture Hotspots You wonât have to ask for problems⌠problems will find you!!
- 11. Outcome (big Picture): The whole process visible Massive learning Critical areas highlighted
- 12. More specifically⌠No scope limitation (paper roll) Exploration of boundaries (External Systems & People) -> The BOTTLENECK is in the picture. -> The CORE DOMAIN is in the picture
- 13. EventStorming blends lean and Theory of constraints into Software Development and Domain-Driven Design
- 14. EventStorming blends lean and Theory of constraints into Software Development and Domain-Driven Design
- 15. EventStorming blends lean and Theory of constraints into Software Development and Domain-Driven Design
- 16. Quick recap of Theory of Constraints Focusing on Bottleneck only
- 17. Bottleneck Main Constraint for a system throughput
- 18. Bottleneck strategies focus on it >>> Subordinate everything else improve it relentlessly small short term improvements will make a difference ⌠while you work on long term improvements too
- 21. Canât find the heart⌠yes, but look at these nailsâŚ
- 22. Improving anything but the bottleneck is close to meaningless Unless you have enough firepower to attack many problems at once..
- 23. The realm of the foggy PO We have plenty of items in the backlog⌠âŚwhose purpose is going to be clear someday
- 24. What Iâll be speaking about and finallyâŚ
- 25. In the last 2 years Been EventStorming in many different situations Different domains large companies & small organisations Established business and startups âŚmany different countries⌠âŚfor different reasonsâŚ
- 26. Biases?? All the possible ones! Iâve seen only scenarios where I was invitedâŚ
- 27. What does a bottleneck look like?
- 28. Digging into the flow
- 29. Unrolling it⌠Business processes expose a repeatable structure
- 31. Letâs look deeper Aggregate Policy / Process Domain Event Command External System
- 32. Letâs look deeper Here is where the system is making decisions: Aggregate Policy / Process Domain Event Command External System
- 33. Letâs look deeper Here is where the system is making decisions: Simple local decision fit into Aggregates Aggregate Policy / Process Domain Event Command External System
- 34. Letâs look deeper Here is where the system is making decisions: Simple local decision fit into Aggregates Aggregate Policy / Process Domain Event Command External System
- 35. Letâs look deeper Here is where the system is making decisions: Simple local decision fit into Aggregates More complex & Reactive decision making fits into policies/Processes Aggregate Policy / Process Domain Event Command External System
- 36. Letâs look deeper Here is where the system is making decisions: Simple local decision fit into Aggregates More complex & Reactive decision making fits into policies/Processes Aggregate Policy / Process Domain Event Command External System
- 37. Letâs look deeper Here is where the system is making decisions: Simple local decision fit into Aggregates More complex & Reactive decision making fits into policies/Processes Aggregate Policy / Process Domain Event Command External System #BusinessProcesses #Transactions #SWArchitecture
- 39. Letâs look deeper Here is where the USER is making decisions: Command/ Decision User/ Actor/ Persona/⌠User Interface
- 40. Letâs look deeper Here is where the USER is making decisions: âŚwhich are based upon real world experience, and the information available on the screen Command/ Decision User/ Actor/ Persona/⌠User Interface
- 41. Letâs look deeper Here is where the USER is making decisions: âŚwhich are based upon real world experience, and the information available on the screen Command/ Decision User/ Actor/ Persona/⌠User Interface #UX #FrontEndDevelopment #UIDesign
- 42. Letâs look deeper Domain Event Read Model User Interface
- 43. Letâs look deeper Domain Event Read Model User Interface Here is were we transform raw data into readable meaningful information
- 44. Letâs look deeper Domain Event Read Model User Interface Here is were we transform raw data into readable meaningful information #BusinessIntelligence #Readability
- 45. And the winner isâŚ
- 46. And the winner is⌠Most of the bottlenecks are here!
- 47. Most of the bottlenecks are here! And the winner isâŚ
- 48. Most of the bottlenecks are here! And the winner isâŚ
- 49. But weâre observing the effect, not the cause
- 51. Annoying anti- patterns in Legacy code
- 52. The Big Ball of Mud
- 54. âThatâs our key classâ
- 55. âThatâs our key classâ Big class with overloaded responsibility in the disputed land between 2 (or more) Bounded Contexts
- 57. Everybody agrees what an order is?
- 58. Everybody agrees what an order is? Of course we do!
- 59. Everybody agrees what an order is? Of course we do! An order is an order!
- 60. Everybody agrees what an order is? Of course we do! An order is an order! And has a customer too!
- 61. Everybody agrees what an order is? Of course we do! An order is an order! And has a customer too! Agreed!
- 62. Perfect recipe: Talk with many people Model Data-First (everybody agrees on nouns) Add some âDogmatic DRY principleâ Repeat
- 63. The Big Ball of Mud
- 64. The Big Ball of Mud I swear it was a monolith last time I checked!!!
- 65. The solution
- 66. The solution A Draft Model with loose integrity constraints
- 67. The solution A Draft Model with loose integrity constraints Constraints implemented as warnings
- 68. The solution A Draft Model with loose integrity constraints Constraints implemented as warnings A Validation barrier (constraints implemented as blockers)
- 69. The solution A Draft Model with loose integrity constraints Constraints implemented as warnings A Validation barrier (constraints implemented as blockers) An Execution Model, with similar data structure, but different behaviour
- 71. Epic win! 1. EventStorming Works! http://www.slideshare.net/ziobrando/why-do-all-my-ddd-apps-look-the-same-vienna-2014
- 72. Epic win! 1. EventStorming Works! 2. I was right 4 years ago!! http://www.slideshare.net/ziobrando/why-do-all-my-ddd-apps-look-the-same-vienna-2014
- 74. Oversimplified, but⌠1) Sketch a logical separation of bounded contexts
- 75. Oversimplified, but⌠1) Sketch a logical separation of bounded contexts 2) Define a local copy of the draft BC: only a few operations required
- 76. Oversimplified, but⌠1) Sketch a logical separation of bounded contexts 2) Define a local copy of the draft BC: only a few operations required 3) âPublishâ the key event, with all the relevant data
- 77. Oversimplified, but⌠1) Sketch a logical separation of bounded contexts 2) Define a local copy of the draft BC: only a few operations required 3) âPublishâ the key event, with all the relevant data 4) Use the published data to instantiate a valid object in the execution BC
- 79. Oversimplification? This one is probably not the only dependency⌠Who knows what else can be there⌠âŚ
- 80. Asymmetries
- 82. Whoever advocates LEGO as an analogy to software development hasnât been doing much LEGO lately
- 83. LEGO Entropy laws Mixing LEGO boxes takes seconds separating LEGO boxes takes ages
- 84. Or, maybe, it just requires some timeâŚ
- 85. LEGO Entropy laws Mixing LEGO boxes takes seconds separating LEGO boxes takes ages Moving a few pieces is reversible⌠âŚuntil you forget Then you loose control and call it âa messâ
- 86. Identify the right bounded contexts as early as possible âŚmodel & linguistic boundaries only⌠In greenfieldâŚ.
- 87. In brownfieldâŚ
- 88. In brownfieldâŚ
- 89. Boy Scout Rule: Leave the campground cleaner than you found it
- 90. The Big Ball of Mud
- 91. The Big Ball of Mud Look, a code smell!!!
- 92. The Big Ball of Mud Look, a code smell!!! Iâve found one too!!
- 93. The Big Ball of Mud Look, a code smell!!! Iâve found one too!! iâve founds dozens of them!!
- 94. The Big Ball of Mud Look, a code smell!!! Iâve found one too!! iâve founds dozens of them!! Letâs rename a few classesâŚ
- 95. ⌠it takes a lot of perseveranceâŚ
- 97. Thereâs asymmetry in payoffs too
- 99. Or, more explicitly⌠Evolution here will have a lot less ripple effects⌠a safer place to improve!
- 100. Or, more explicitly⌠Evolution here will have a lot less ripple effects⌠a safer place to improve! And here situation is better tooâŚ
- 101. Or, more explicitly⌠Evolution here will have a lot less ripple effects⌠a safer place to improve! And here situation is better too⌠1) we are reducing the number of people related to a model
- 102. Or, more explicitly⌠Evolution here will have a lot less ripple effects⌠a safer place to improve! And here situation is better too⌠1) we are reducing the number of people related to a model 2) we are driven by business, not by smells
- 103. Smell is a good driver if you have no idea where the money is
- 104. Refactoring danger zone In code: Dependencies Compiler Tests In the database A lot harder to know who will be affected
- 105. Itâs not the cleanup, itâs the associated risk It takes only one violation to get into âWe have no ideaâ zone
- 107. Unclear payoff -> Risk avoidance clear payoff -> Risk Management
- 108. What happens when you discover the bottleneck? âŚmaybe not what you expectedâŚ
- 110. found the bottleneck, now what?
- 111. found the bottleneck, now what? Write some documentation?
- 112. found the bottleneck, now what? Write some documentation? Call another meeting?
- 113. found the bottleneck, now what? Write some documentation? Call another meeting? Letâs think about it!
- 114. found the bottleneck, now what? Write some documentation? Call another meeting? Letâs think about it!
- 115. found the bottleneck, now what? Write some documentation? Call another meeting? Letâs think about it! ⌠write a prototype?
- 116. found the bottleneck, now what? Write some documentation? Call another meeting? Letâs think about it! ⌠write a prototype? implementing a solution?
- 118. found the bottleneck, now what?
- 119. found the bottleneck, now what? Write some documentation?
- 120. found the bottleneck, now what? Write some documentation? Call another meeting?
- 121. found the bottleneck, now what? Write some documentation? Call another meeting? Letâs think about it!
- 122. found the bottleneck, now what? Write some documentation? Call another meeting? Letâs think about it!
- 123. found the bottleneck, now what? Write some documentation? Call another meeting? Letâs think about it! ⌠write a prototype?
- 124. #protip schedule a free day after an EventStorming
- 125. But the scary thing isâŚ
- 126. Some companies attacked the problem, some didn't
- 129. The Dungeon Master https://medium.com/@ziobrando/the-rise-and-fall-of-the-dungeon-master-c2d511eed12f#.rtbyh8uid (yes I am aware that this one is the Star Wars emperor)
- 131. A few lessons rationality doesnât explain everything habits, unconscious, weâre not machines. Passion and Pride can become attachment LEARNING canât be skipped or accelerated: sometimes things need to be discovered again
- 133. Strong emotions How would what if three lines of your code created a million euro hole? Code persist, but the context is gone Itâs how people feel, not how they should that drives some behaviour.
- 134. Are people feeling safe in changing?
- 135. Do they have options? Legacy is whatever youâre afraid to change
- 136. Culture is a better asset than code Safety, Experiments, âŚand component size too!
- 137. Conclusions âŚmaybe not what you expectedâŚ
- 138. Enables cross- perspective conversation Software design, User Experience, Business Modelling, Lean etc.
- 139. Can lead you to unexpected places apparently, you cannot choose what your problem is
- 140. Stop building monoliths I know you wonât ⌠itâs for slideshare. ;-)
- 141. I wasnât explicit enough
- 142. Data Centric applications are damaging your business I wasnât explicit enough
- 143. Data Centric applications are damaging your business The only thing that keeps companies in the market is that competitors are in the same situation⌠I wasnât explicit enough
- 144. Thereâs a recurring way out
- 145. And a recurring problem too #DungeonMaster
- 146. Stop ignoring the culture We are not rational human beings 24H/day Emotions are driving technical choices more than you think
- 147. Thank you!
- 148. References ⢠www.eventstorming.com ⢠EventStormers on Google+ ⢠https://plus.google.com/u/0/communities/ 113258571348605620818 ⢠LeanPub book in progress: ⢠http://leanpub.com/introducing_eventstorming ⢠Blog: ⢠https://medium.com/@ziobrando ⢠http://ziobrando.blogspot.com ⢠Twitter: @ziobrando ⢠Trainings & Workshop facilitation: ⢠http://www.avanscoperta.it