Initiation to Code: A Roadmap for New Developers and Their Mentors

Starting your first programming job can be a daunting experience. How do you progress from basic CRUD apps to competent work on a complex application? Meanwhile, training inexperienced developers can feel like a goal with no roadmap. How can you help a newbie when you can’t even remember what it was like to be in his shoes? Drawing from a wide range of topics including cognitive psychology, mysticism, and alligators, this talk will address both sides of the technical onboarding challenge, providing thorough and structured advice for both entry-level developers and those who mentor them. Many of the tips on skill acquisition and learning should be helpful to those in more advanced positions as well.

Transcript

1. Alice Mottola Initiation to Code A Roadmap for New Developers and Their Mentors
4. Baby gators = junior developers Big gators = confusion and anxiety Swamp tour guide = mentor
6. “junior software”: 8,874 “senior software”: 89,451
7. “junior software”: 16,069 “senior software”: 116,695
8. “junior software”: 1,848 “senior software”: 12,042
9. “junior software”: 1,679 “senior software”: 1,797
10. “junior software”: 1,679 “senior software”: 1,797 That’s different...
11. Junior developers are good for your team. • Challenge more experienced engineers to clarify and refine their knowledge • Practicing communication makes your team stronger • People do their best work when they’re treading water ◦ Best case scenario: you’re challenging everyone, all the time ◦ Junior devs make that easier to achieve
12. But training is so haaaaard...
13. But training is so haaaaard... Nah, it doesn’t have to be.
14. Structure • Three phases based on the noob’s comfort level • Each phase covers advice for both the mentor and the noob • No time estimates — everyone is different
15. PHASE 1: Learning to Swim • You know some basics. • You’ve never gone swimming in a large code base before. • You have no experience working on a team making software that people actually use.
16. PHASE 1: For Mentors
17. PHASE 1: For Mentors • Best case scenario: assign the noob a dedicated mentor • This person: ◦ Is a good communicator ◦ Actively budgets time for the noob • Significantly lowered onboarding time at New Relic • Even if you’re not a dedicated mentor, you can still use the advice from this talk
18. Communication: A Template • Provide a framework for learning without spoon feeding all the answers • Think of the template method pattern...
19. class Developer def do_stuff debug learn_pattern learn_quirk_of_design end def learn_quirk_of_design # teach your noob! end end Communication: A Template
20. class Developer def do_stuff debug learn_pattern learn_quirk_of_design end def learn_quirk_of_design # teach your noob! end end Communication: A Template class JuniorDeveloper < Developer def debug # fix that bug! end def learn_pattern # study that pattern! end end
21. class Developer def do_stuff debug learn_pattern learn_quirk_of_design end def learn_quirk_of_design # teach your noob! end end Communication: A Template class JuniorDeveloper < Developer def debug # fix that bug! end def learn_pattern # study that pattern! end end
22. Communication: A Template • You fill in: anything unusual or unique to your application • They fill in: pretty much everything else
23. Communication: A Template • Don’t assume knowledge • To get a sense of your noob’s initial level, quiz him a bit ◦ Don’t ask “Do you understand nested resources?” ◦ Instead, ask “How do nested resources work?” • If he doesn’t know some fundamental things, have him go read about them • If he has it covered, start him on some simple assignments
24. Communication: A Template • Don’t restrict your guidance to the technical realm ◦ Introduce the team process ◦ Tell your noob who is working on what ◦ This is easy to forget, but it helps your noob a lot! • Be honest about your own limitations ◦ If you’re busy or you run out of patience, let him know politely ◦ If you’re confused, say so
25. IMPORTANT. Pair programming is critical at this stage. It’s not optional. It’s mandatory.
26. PHASE 1: For Noobs
27. Metacognition.
28. Metacognition!!
29. Metacognition is your brain’s skeleton key. • Active control over the cognitive processes involved in learning • Tightening the feedback loop with yourself • Correlated with general intelligence
30. Metacognition is your brain’s skeleton key. • Plan an approach • Monitor the process • Reflect on the outcome
31. Set explicit goals and expectations. “Plans are useless, but planning is indispensable.” – Dwight D. Eisenhower
32. Set explicit goals and expectations. • It doesn’t matter if your goal changes in the middle; in fact, that’s great! • Not just for big things; for anything ◦ Reading documentation for a gem you’re about to use ◦ Writing a blog post ◦ Going into a meeting • Write it down
33. Empower yourself with curiosity.
34. Empower yourself with curiosity. Pitfall #1: Feeling powerless
35. Empower yourself with curiosity. • Stage I: Curiosity • Stage II: Control • Stage III: Conscious intent – Wallace Wang’s outline of the hacker mentality, from Steal This Computer Book 4.0
36. Empower yourself with curiosity. Pitfall #2: Going mad with (false) power
37. Empower yourself with curiosity. Feeling powerless Going mad with (false) power Your version of the hacker mentality
38. Memory Palaces
39. Memory Palaces • Effective technique for encoding information into long- term memory • Oldest known mnemonic strategy ◦ Described by Cicero in De Oratore
40. IMAGINE YOUR CHILDHOOD HOME...
41. Memory Palaces • A walk through a familiar place • Add imagery representing what you want to memorize • Make the images as outrageous and weird as possible • You can also try using imaginary locations with characteristics that represent the ideas
42. Help Your Team with This One Weird Trick!
43. Help Your Team with This One Weird Trick! USE THE PRODUCT.
44. Phase 1 Summary • Provide a framework for skill acquisition without spoon-feeding • Be honest about your own limitations • Pair program Mentor
45. Phase 1 Summary • Provide a framework for skill acquisition without spoon-feeding • Be honest about your own limitations • Pair program Mentor • METACOGNITION • blah blah other stuff Noob
46. PHASE 2: Touring the Swamp • The path is clearer and the stumbling blocks are well- defined. • You can work independently on small things, but you still need guidance often. • You usually have a rough idea of what to do next, but less often how to do it.
47. PHASE 2: For Mentors
48. Help your noob stay challenged. • This phase is all about improvisation • DON’T just assign whatever happens to come up
49. Help your noob stay challenged. • Your objective: Keep your noob’s head just barely above the water • Much like decorators add functionality to an object at runtime… ◦ You must improvise to add new functionality to your noob
50. Help your noob stay challenged. • Pay attention to what your noob has worked on and favor assignments that are harder or dissimilar ◦ If you can’t remember what your noob has done, it’s okay to ask
51. Help your noob stay challenged. • Train your noob to look at her code with a pessimistic eye ◦ What could go wrong? ◦ What assumptions have we made that could sometimes be false? ◦ What would a null input do? ◦ Does it open any attack vectors?
52. Encourage. • Could be as simple as saying “Good work” ◦ This is very easy to forget! But don’t forget! • Tell your noob when his code has been deployed ◦ Shows the noob his work matters ◦ Familiarizes him with the process ◦ “No, your code is not just being swallowed by a dark void.”
53. Encourage. • The next tip is difficult to apply gracefully, so please take care...
54. Encourage. • The next tip is difficult to apply gracefully, so please take care... • Don’t let your noob be a wuss
55. PHASE 2: For Noobs
56. Be a code entomologist. • You’ll probably be debugging a lot at this stage • Whenever you fix a bug, think about what you got hung up on • Develop a taxonomy ◦ Generalize your bug as part of a category • “How could I have figured this out faster?”
57. Be a code entomologist. • The SLT (Stupid Little Thing) ◦ The error is in syntax or spelling ◦ Smells like: complete understanding blended with mystifyingly broken code
58. Be a code entomologist. • The SLT (Stupid Little Thing) ◦ The error is in syntax or spelling ◦ Smells like: complete understanding blended with mystifyingly broken code • Framework gotcha ◦ A component of the framework you’re using works in a way you didn’t expect ◦ Smells like: Mild confusion with a hint of changelog
59. Be a code entomologist. • The principle here is: ◦ Notice when you’re deep in a rabbit hole ◦ Take a step back ◦ Compare to previously encountered bugs
60. Have opinions, even if they’re wrong. • If you change your mind later, you’ll remember it forever • Emotional connection strengthens memory retention • Deeper understanding of the tools you use ◦ Their philosophy, what they do and don’t do • Open opportunities for great conversations with people who know more than you
61. On that note, review others’ code. • Even if you don’t feel qualified • Thoughtful exchanges are not a waste of time • The reviewee will most likely appreciate it
62. Find an independent project. • This is the ideal time for a self-driven project • A personal project is okay, but richer experience if it’s for your company or team ◦ Lots of potential collaborators ◦ Real usability testing! ◦ Accountability is almost as good as a deadline • Using the product (cf. Phase 1) can provide ideas
63. Find an independent project. • Otherwise, try an internal tool • What would help your team be more effective? ◦ Faster? ◦ More knowledgeable? ◦ Easier to communicate?
64. Phase 2 Summary • Objective: maintain the right challenge level • Tactics: ◦ Improvise ◦ Add depth • Don’t forget to encourage Mentor • Catalogue what you learn from bugs • Invite opportunities to be wrong • Start working on an independent project Noob
65. PHASE 3: Swimming Free • You still have a lot to learn, but you can work independently most of the time. • Progress is mostly self-driven. • Pairing sessions feel more collaborative than instructive.
66. PHASE 3: For Mentors
67. How does a noob get from Phase 2 to Phase 3?
68. Initiation
70. What makes an initiation? • Would not be trivial for someone at a higher level • Demands communication with senior engineers • Provides demonstrable (preferably quantifiable) value
71. My rite of passage...
72. My rite of passage... “There are only two hard things in Computer Science: cache invalidation and naming things.”
73. My rite of passage... Rails.cache.delete_matched("views*")
74. My rite of passage... Rails.cache.delete_matched("views*") BAD!
75. My rite of passage... • Build predictable keys based on a set of attributes (sort type, request type, account ID) • When cache is busted for a given account, build all possible keys and delete each ◦ Much faster than a wildcard match!
76. My rite of passage... • Build predictable keys based on a set of attributes (sort type, request type, account ID) • When cache is busted for a given account, build all possible keys and delete each ◦ Much faster than a wildcard match! But what if we add a new sort type? (The code shouldn’t have to change.)
77. My rite of passage... • Store any encountered value for sort type in Redis • “But I thought we were reducing load on Redis!” Yup. • So add an intermediate layer of caching at the class level ◦ Thread-safe key-value store in controller ◦ If a value exists in the class-level cache, it’s already in Redis, so don’t store it again
78. My rite of passage... • Store any encountered value for sort type in Redis • “But I thought we were reducing load on Redis!” Yup. • So add an intermediate layer of caching at the class level ◦ Thread-safe key-value store in controller ◦ If a value exists in the class-level cache, it’s already in Redis, so don’t store it again Oh, and test all this.
79. My rite of passage...
80. What makes an initiation? • You don’t have to tell your noob it’s an initiation • In fact, you probably shouldn’t
81. Post-Initiation • Keep up the same communication style as before — but a lot more hands-off • Introduce the noob to new ideas ◦ Elaborate or suggest resources on topics she’s interested in
82. PHASE 3: For Noobs
83. Invite your own initiations. • Keep working on your independent project(s) ◦ If you haven’t started coding in your free time yet, do it now ◦ You might be surprised how easy it is • This routine works for me: ◦ Directly after getting home, 15–20 minutes of meditation ◦ Then, 45 minutes of “creative time” ◦ Creative time can be writing, coding, or drawing ◦ NOT reading
84. Invite your own initiations. • Read texts that are over your head.
85. Invite your own initiations. • Read texts that are over your head. ◦ Seriously, OVER YOUR HEAD!
88. (Disclaimer: Sandi rules and POODR is a great book!) (But you should be reading stuff that makes your head threaten to pop off.)
89. Challenging texts in Ruby/Rails
90. Challenging texts in Ruby/Rails
91. Challenging texts in Ruby/Rails
92. Challenging texts in Ruby/Rails tenderlovemaking.com
93. Invite your own initiations. • Go to conferences ◦ Especially hacker conferences! ◦ Try some events that are outside the scope of what you do at work • Learn a new language, especially in an unfamiliar paradigm ◦ C or a purely functional language are good choices
94. Teach yourself design patterns. • The most important thing you don’t learn in boot camp • High ROI for your time ◦ Easy to understand ◦ Extremely useful ◦ Makes you a better person on the whole, really
95. Be a mentor.
96. Phase 3 Summary • Set up an initiation • Then broaden the scope Mentor Noob • Invite further initiations • Learn design patterns • Be a mentor
97. What did we learn? 1. Develop metacognition 1. Communicate 2. Apply metacognition 3. Invite challenges 2. Challenge 3. Initiate NOOB MENTOR
98. Your homework:
99. “A true initiation never ends.” —- Robert Anton Wilson
100. References / Influences • Chuck Lauer Vose, “Building kick-ass internal education programs,” RailsConf 2014. https: //www.youtube.com/watch?v=LPZmNfhPPOs • U.S. Bureau of Labor Statistics. http://www.bls.gov/ooh/computer-and-information- technology/software-developers.htm • Vanderbilt University Center for Teaching, “Metacognition.” http://cft.vanderbilt.edu/guides- sub-pages/metacognition/ • Wallace Wang, Steal This Computer Book 4.0 • Jose Valim, Crafting Rails Applications • Pat Shaughnessy, Ruby Under a Microscope • Aaron Patterson, “Tenderlovemaking.” http://tenderlovemaking.com/ • Eric Freeman, Elisabeth Freeman, Kathy Sierra, Bert Bates, Head First Design Patterns • Grant Morrison, The Invisibles • Robert Anton Wilson, Masks of the Illuminati • Cajun Pride Swamp Tours, LaPlace, LA
101. Thank You! Questions? @ichthala ichthala@gmail.com github.com/ichthala