These are notes for the Apprenticeship Programme. Distributed under Creative Commons Attribution Sharealike International 4.0
This programme intends to embed the same baseline knowledge into a new person, to avoid discovering holes in their knowledge later.
This programme needs a one-to-one pairing of apprentice with a senior developer. It must be implemented within existing team structure, already working on some project. The reason for such requirements is to present the apprentice with real-life problems. Because of the above it works best with long-term assignements, where you invest in a person and your team dynamics up front to get the work quality you need.
Recommended timeframe for covering the topics below is 3 months, however can be shortened/made longer depending on the progress of the apprentice.
Tested on several mentor-apprentice pairs, always with local pairing. No remote experience here, would probably need adjustment for remote settings. Have such experience - please contribute !
If you have experience being an apprentice or a mentor - feel free to add a pull request here, however small !
Don't want to create a pull request ? Starting an apprenticeship programme and unsure how it should be run in practice ? World seems bleak and stripped of joy ? Contact me: email/jabber/xmpp: cyplo@cyplo.net
- for the apprentice to be able to learn on their own
- baselining knowledge - this allows other, equally trained, people to expect knowledge of the subjects below when working with the apprentice
- for the apprentice to feel comfortable when asking questions
- allows the apprentice to formulate a plan of approaching a problem on their own - do not hand over a complete plan of actions
Before apprentice starts:
- decide on the mentor
- make sure that everyone understands that we're investing the time now to iron out all of the issues today - this means that the whole team is taking a performance hit by accepting the apprentice, and they do want to do it anyway.
- make sure to announce when the apprentice starts to everyone involved.
Starting
- look through this document
- look through other materials available
- discuss your experience, come up with a plan for first week together
- after the week come up with the plan for the next week and so on every week
- treat this document as a rough sketch of a overarching plan/list of goals
- revise the progress on the bigger picture every week
What's where
- office supplies
Who's who
- show the apprentice how to locate subject-matter experts using in-house tools
- introduce people when working on particular problems [X might know about this, let us ask her !]
Office tech
- email setup
- email groups
- IM setup + testing
- VPN setup + tests
- WiFi
Company network
- intranet
- distribution lists
- Wiki
- Task management
Project infrastructure.
- Build servers
Knowledge sources
- schedule slots for reading - make Friday the day of learning
- Fowler's "Refactoring"
- RSS feeds [share people's feeds - see .opml files in the repo]
Noting stuff down
- introduce to 'note everything, sort later' aka 'post mortems and time management for free' approach
Learn to learn
- review mistakes and sucesses every Friday
- open discussions with your peers, and not being afraid to ask stupid questions
commandline:
- bash/zsh [cygwin if Windoze]
- pipes, redirects
- when in doubt use the simpliest, most brute force tools possible, i.e. grep and find. introduce grep as a tool to check for mistakes during refactorings, etc
- involve grep and find for finding places in legacy code
SCM
- share configs, especially diff tool setups
- spend some real time explaining basics of centralized vs distributed solutions
- good branching model reaching
- interactive git tutorial
- how to revert anything in git
- always look at diffs before commiting
project setup -> ability to deploy
Algorithms:
- RSA
- hashes and dictionaries/hash maps
Memory management:
- mapping between native memory and JVM/CLR memory
- GC
- heap
- stack
Computer bootup process
Networking:
- ssl/tls
- C10K problem, implement singlethreaded low level server on raw sockets
IPC
Multitasking:
- OS-level threads vs green threads
- async io
- immutability of your data and multithreading - functional programming vs object oriented
Working with code
- codetiquette - remember that in most cases you are not the only one working on this code
- code standards - the more boring the better, the more uniform the better
- ability to read code
- working with legacy code [freeze with tests + cut out approach]
- negate condition -> decrease indentation
- guard clauses at the beginning
- ifs/switch -> dictionary
- when to refactor ?
- symmetry broken
- continuous refactoring - improving your codebase rather than regressing it
- error and exception handling
- do not use exceptions for flow control
- talk about the difference in execution paths, how exceptions affect stack etc
- introduce monadic error handling
- care about what you can handle, pass all other stuff upwards
- always reproduce a bug with fully automated test first
- good naming: what it does, not how it does it
- introduction to TDD
- inner TDD loop vs outer, acceptance-tests driven TDD loop
- spikes in TDD
META
- Recommend changes to the apprenticeship programme
