Science-based Development

Development, like science, is a messy endeavor. It’s near impossible to control all the variables. Testing code is a noble goal, but it’s very easy to test the wrong thing. Many experienced scientists are fooled by their senses and biases. To account for that they crafted and refined a Scientific Method. That method has been stress tested by centuries of experimentation, discovery and peer-review. It probably wouldn’t hurt to try to apply it to development and see what it can offer.

Transcript

1. The Method Science-based Development
2. process.
3. Ain’t nobody got time for dat!
4. You’re... doing it wrong?
5. Science can help!
6. It can make you be er.
7. I dare ya! prove me wrong.
9. more hair hair not where I live photo non contractuelle @olivierlacan
12. Things Science cares about: - me: nope - you: nope - truth: yep
13. Science Use Do Things To
14. Do Things With Science
15. is a method Science
16. is ok with wrong Science
17. There was an elderly professor who had been passionately keen on a particular theory for a number of years. One day an American researcher came and u erly disproved our old man's hypothesis. The old man strode to the front, shook his hand and said, “My dear fellow, I wish to thank you, I have been wrong these fi een years”. And we all clapped our hands raw. That was the scientific ideal, of somebody who had a lot invested, a lifetime almost invested in a theory, and he was rejoicing that he had been shown wrong and that scientific truth had been advanced. Richard Dawkins in The Root of All Evil? Story Time
18. doesn’t trust us Science
19. liar
20. pushed us forward Science
22. Scientific Method The
23. 1 Inquiry
24. curiosity it totally didn’t kill the cat, you guys!
27. “I have no special talents. I am only passionately curious.” Albert Einstein
28. seek knowledge
29. find problems
30. distrust authority
31. Kary Mullis, biochemist
32. “When I got the idea one night that I could amplify DNA (...), and that I could make lots of copies of some little piece of DNA, the thinking for that was about 20 minutes. (...) I did talk to people about it but if I’d listened to what I heard from all my friends who were molecular biologists I would have abandoned it.”
33. too young don’t let ideas die they might be crazy enough to work
34. 2 Hypothesis
35. beware of assumptions
36. this is what our users want we need this feature right now without a Like button, we’re nothing
37. ask a question
38. ask the right question
39. it can be broad beware
40. it can be specific beware too
41. it must be testable
42. it must be falsifiable able to be proven wrong
43. it must predict “if Y, then X should happen”
44. using AJAX here should improve our sign up rate by at least 10% falsifiable statement
45. - down 1% sign ups: - up 0% - up 5% - up 10% - up 20% hypothesis falsified hypothesis supported
46. 3 Experiment
47. reality where ideas come to die
48. protocol experimental
49. isolate variables
50. enforce controls what was the real cause?
51. check the null hypothesis try to prove yourself wrong
52. check your bias at the door it’s easy to fool yourself
53. participant doesn’t know experimenter knows single-blind
54. double-blind participant doesn’t know experimenter doesn’t know
55. triple-blind participant doesn’t know experimenter doesn’t know result analysts don’t know
56. who’s who? you code participant experimenter user
57. A/B Testing is OK. provided you don’t forget about: C D E F G H I J K L M N O P Q R S T U V W X Y Z
58. Actual Testing is OK too.
59. describe User do let(:user) { Factory.build(:user) } subject { user } it { should have_many(:achievements) } it { should have_many(:courses) } it { should have_many(:orders) } it { should have_many(:prizes) } it { should have_many(:videos) } end predictions
60. context '#complete_course!' do it "increases count of completed courses" do expect { user.complete_course!(course) }.to change { user.courses.completed.count }.by(1) end end
61. Rinse. Repeat. Forever. (continuous integration)
62. 4 Analysis
63. What happened? can other questions be asked?
64. Failure is OK. improve the hypothesis or start from scratch
65. Reproduction was the result isolated?
66. Publication share your results and your tools!
67. Science Inject some in your DEVELOPMENT
69. you’ll pay closer attention to worst case scenario: THINKING & ANALYZING DEVELOPING OBSERVING
70. bye now.
71. @olivierlacan feedback, rebuttals, experiences, etc.