-
Notifications
You must be signed in to change notification settings - Fork 87
Retrospectives
Arnaud Bailly edited this page Jul 15, 2021
·
3 revisions
See Miro board for original material. Some more notes below from discussion about each issue.
- 😄 / 😦 No Jira:
- Not feeling confident with making promises, not confident with planning aspects
- lots of gut feeling for estimates
- Jira is not synonym of structure, we do have a structure => there's a way to get the good things (projections) w/o the bad things (overweight)
- Miro does not try to promise certainty/exactness
- => have a way to give estimates?
- 😄 Sebastian's Leadership
- Doing great
- 😄 Testing: Discussed a lot about it and bit of a friction but feel confident we are on the same page
- Keep pressing on not only doing unit tests, helps a lot fleshing the implementation
- Testing gives confidence in the long run, saves time. Everyone is convinced we should do it.
- 😄 Good communication w/in the team
- disagreement does not lead to resentment
- missing it when someone is not there, feeling less energy with one less people
- pairing/ensemble programming feels more productive with 2 people?
- ensemble with 3 adds more insights, content, feels less efficient but more effective
- 😄 Pairing/mobbing:
- => obey the timer more strictly
- => decrease the length of cycles
- sometimes doing and erasing later it is faster than discussing
- 😦 source dependency management is a pain
- diamond dependencies with conflicts through various pieces of the code
- nix does not help there -> additional roadblock in our setting
- use of haskell.nix is clunky?
- have a trimmed down nix-shell which only builds system dependencies and use plain cabal to build all haskell deps
- options seem to be: keep the nix stuff and train ourselves or switch entirely another tool?
- haskell.nix does not really work with stack because no one is using it
- => spend some time next week deciding how to move forward
- 😄 Miro:
- everything is in the right place
- 😄 No Project manager
- we are still doing project management
- there is not a team depending on us, we depend on other teams, synchronizing happens "informally" in the engineering meeting or through conversations with people
- Arbitrary Bureaucratic PM Work -> just filling out forms, just more work
- project managers are mostly doing reporting anyway
- 😄 Demo:
- no glitch, very confident to add something and play with it live
- 😕 Tail research:
- good to be involved but very unbounded work, could drift off
- things are not clear on the amount of work needed
- how to manage expectations in a better way? we are doing dev and also helping research
- => kanban with dedicated 1/6th slot for research work
- 😦 No users yet:
- have something out that people could start using or give us feedback on
- => build more documentation!
- 😕 Plutus:
- feels somewhat overwhelming yet feel better than a month ago
- slow, hard to get feedback, errors are hard to decrypt
- => keep isolating plutus specific code
- 😄 property tests:
- use it more often, even where there's a single test
- we sometimes have tests about some transaction => transform into PBT
We do a talking-stick style retrospective where everyone talks uninterrupted then hand-over the (virtual) talking stick to their "neighbour" until everyone passes.
- AB: appreciate the debates and the ways we are doing thing, eager to start something
- SN: doing things on the side is good, appreciate it. Sometimes struggling a bit with discussions, not talking about the same thing, different context
- Like we do this discussions, have a better result in the end, discussions further what we have and we do
- Getting our hands on something, get back coding and show something, have a walking skeleton to support discussion but discussions were needed
- We know stuff, we can ignore what we know and revisit all of this once we have a prototype
- MB:
- Getting started to do something
- Many discussions these past days, there are tradeoffs, not much participation and it was deliberate, "want to go with the flow"
- Past 3 years arguing on what to do and how to do, we can try ideas and check there are goods, we'll have plenty of time to rewrite code
- We need someone to take the decision?
- We have done many small things, we struggle to find agreements on some things though
- Happy with Plutus stuff, not because of plutus, having a goal and be able to do something, feeling of accomplishment
- AB:
- debating without concrete objectives / arguments (bikeshedding designs?) is something we should try to avoid
- need structure, avoid the "tyranny of structureless"
- don't want to do PRs and Code reviews
- SN
- MB:
- having one person do all things is fraught with perils
- not only about trust
- getting used to PP, better this week and last week thanks to being able to focus on Hydra <= having a single day on wallet
- drawings are very concrete, but we have many (too many?) of them, need a way to make decision to have a "first version" of drawing
- SN:
- great that Matthias can focus more, important to have MB around
- schedule is vague, objectives are fuzzy, we should come together to have more specific objectives
- reflecting on milestone objectives: making sure we have an architecture for the product
- 2nd objective: scope, finding a sweetspot between constraints and requirements, limitations
- 3rd goal: prototype, getting closer to that goal, but it's only a prototype not set in stone, there are other ways
- bikeshedding is not an issue if you have a time clock running!
- open, but wouldn't mind with a traditional workflow
- AB:
- tried mob and it worked
- about taking decisions: we should have a project charter in which we define our decision process
- MB:
- want to avoid bikeshedding and lengthy PRs
- small team is good, you don't go faster adding more people
- SN:
- like the fact we are 3 people; but at some point we might need to hire more engineers; maybe after the prototype; what about Julian Ospald?
- risk/story-map was a good decision, get the landscape mapped, find "skeletons buried"
- when do we need structure? having a backlog which is guiding the effort pursued as an ensemble
- right now it does feel too adhoc, mixing uneasiness with relating PRs with US -> help have a better sleep
- "how can we get Matthias get a better sleep?"
- how do we feel about having Jira and use that for a prioritized backlog of our work?
- AB:
- I appreciate on that team
- backlog is totally fine
- it's good and important to set goals for ensemble sessions, even for each keyboard-passing-round
- MB:
- make something out of the team
- how do we measure progress -> Team Charter
- right now, by the number of "Green stickies" on the roadmap
- define a way to define progress? milestones, goals, how capable is our Hydra system
- SN:
- good question about progress, will change across phases
- progress would mean we have a demo -> nobody cares about protocol/network, we need something that can give an idea of what the product can be
- really likes to be here