Team best practices
This document aims to address contibutors best practices of the following repositories:
This document is not in its final version, a team meeting which aim to address new/old best practices, feedback, workflows, all kind of issues related to how the team work together occurs every 2 weeks.
This document link to other specialised best practices (like coding best practices).
### 1) Team meetings:
Daily standup (1pm CET - 11am GMT) for taking care of the current issues.
A regular standup - each Tuesday (3pm CET - 1pm GMT) - which aim to
- Update every contributor on what others are doing.
- Update the prioritised issues / PRs list.
- Address little issues (possibly related to the current ongoing milestone).
- High level demo, explanation about specific points of the codebase or Ethereum related things.
A milestone standup - scheduled before the beginning of each milestone, roughly on a monthly basis - which aim to define what will be included in the next milestone and who will work on what. This standup also help to set a clear long term vision.
A retrospective standup - after each releases - which aim to talk about best practices in general: what is good, what is bad, how we can improve workflows.
A tour standup - Just after a release or whenever it is needed - which aim to demo, explain in details features, bug fixes or any part of the codebase.
### 2) Group meetings:
- When a story / bug fix is divided in parts, there should be a kickstart meeting with all the developers involved, so that all the devs have an good overview / understanding on:
- How the story fits into the Ethereum tech.
- How the backend (if any) works / will work (could be a smart contract).
- How the frontend works / will work.
- What is the general vision of the UX design for this particular story.
Later progress and discussion is updated directly on the issue or pull request (Github).
Before starting coding, we should ensure all devs / contributors are aware of:
Story / Bug fix
- Prioritised list of PRs / issues are tracked in a Github Project, Remix IDE issues are managed by a prioritized backlog.
- Every story can be executed by a single developer or a group of 2 or more developers (depending on the size and complexity)
- Each dev should take the part he/she feels the most confortable with.
- Later progress and discussion is updated directly on the issue or pull request (github).
- When a developer or team decides on the story they want to work on (at the start of milestone for instance), they assign themselves to the issue.
- Documentation update should be done together with the story, or an issue with the label "documentation" has to be created.
### 1) PR Creator:
- It is recommended to use the emoji responses to signal agreement or that you've seen a comment and will address it rather than replying. This reduces github inbox spam.
- Mark unfinished pull requests with the
Work in Progress label
- Large pull requests (above 200-400 lines of code changed) cannot be effectively reviewed and should be split into smaller pieces.
- Code should comply to the
- You should not expect complete review on a pull request which is not passing CI.
- You can obviously ask for feedback on your approach.
- You should assign a reviewer(s).
- Pull requests should be used as a reference to update coding best practices whenever it is needed.
### 2) Review:
- Everyone is free to review any pull request.
- You should add the label "change requested" or "accepted".
When reviewing people's code consider the following two comments.
I don't like the name of this function.
What do you think about changing the name of this function to ....
Your feedback will often be better received if you pose it in the form of a question.
Pull request should be reviewed to comply to coding best practices.
You should take the responsability of the PR you are reviewing.
You should make sure the app is viable after the PR is being merged.
You should make sure the PR is correctly tested (e2e tests, unit tests)
Ideally You should have enough knowledge to be able to fix related bugs.
### 3) Merge:
- Merging is possible after Review and Tests are ok and when the PR is approved.
- After a merge, it is highly recommended to check the new code in
- A milestone should only contain items we are sure to finish.
- The end of a milestone trigger a new release.
- Milestone items and duration should take in account time spent in bugs fixing and support.
- The team should commit to the milestone duration.
- If a dev finish early he/she can help other to push remaining tasks.
- If a dev finish early he/she can work on specifying / integrating the next milestone.
- A milestone duration is fixed at the start of the milestone (but should better not exceed 1 month).
- Progress and issues regarding a milestone are discussed on regular standups.
### 1) Process:
- Should be documented and updated.
- A new release is triggered:
- after an important bug fix
- at the end of a milestone
- We can release an
m.m.m-alpha.x (whenever we need to release and for whatever reasons) being in between a feature / bug fix completion.
- We release an
m.m.x whenever there is a bug fix.
- We release an
m.x.0 whenever there is a new feature.
- We release an
x.0.0 after each milestone we consider being an important progress.
- We release an
x.0.0 if there's an API breaking change in one of our libraries.
- After a new release we should stay in alert for possible regression and better not release Friday at 5pm :)
### 2) Community:
- Before the official release, we should select a group of power users and invite them to test and give feedbacks.
- Users need to know upfront a new release is coming and we should prepare them for it by showcasing some new features they can expect and when it will happen (fixed date, published at least 1 week in advance).
- Whenever we have a new release we have to communicate this efficiently (twitter, reddit, ...).
- A critical bug should get the label
Blocker, and every effort should be put to fix it.
- Addressing a non critical and non planned bug can be done:
- After having notified in the
remix-dev channel if the bug does not involves UX or public API changes.
- After a dev meeting (e.g the regular standup) if the bug involves any UX or public API changes.
- We should all keep an eye on the public non dev channel and file user feedback.
- The documentation is done / updated just after the feature / release in a team effort.
- Documentation work is filable as a github issue.
- It is encouraged to find and link associated doc produced by the community (blog posts, videos, tutorials, ...)
Coding best practices
Best practices to address later
- Find a good calendar we can use to schedule meetings.
- Retrospective meeting: Someone can be responsible to record the findings (of the retrospective meeting) so we can review them in the following meeting and see if they were addressed or not, and if not discuss why and take new actions. If the issue was addressed then it can be discarded.
- pair programming meetings
- All the contributors of the feature should be able to merge.
- Review should have 3 phases (1. code should work, 2. code should fit with current best practices, 3. update documentation)
- We should have a dedicated person to only do browser frontend tests and make it easy for others to run them
- We release an
x.0.0 if there's a fundamental change in our UX design, which means users will need to readapt the way they use the app
- after a week finishes, we publish/release a new version as remix-beta.ethereum.org and inform users so early adopters can test. after another week, when then next finished work is released as remix-beta.ethereum.org, the previous one becomes remix.ethereum.org and all users can start using it
- a bot to automatically notify users about upcoming features on all channels whenever remix-beta.ethereum.org is updated
in case it's a major version increase - this announcement should be specially marked so ppl can check early instead of being confronted with drastic changes when remix.ethereum.org updates
Setting up a "bug" time where we each take a bug for which:
We set up a special day where we address all the necessary documentation work in a team effort.
Change to markdown or gitbook
- Remix channel - we rotate and each day one or two are responsible for support (It would also be important to know for those of us who are contracted, how this can be billed.). If they don't know the answer they ping team member who they think could answer. That team member checks the support chat that day only if she/he is mentioned.
- We should have a FAQ where basic answer are written down so we can drop the link regularly in the channel
- People should be pointed to the appropriate
best practice.md files by team members
code best practices: