---
tags: weekly-meeting
---
# 2023-06-19 types team retro
Retrospective Survey Summary
Survey data:
- 7 responses
- 4 from types team
Positive Aspects:
1. Sense of Momentum and Energy:
The survey indicated a positive sense of momentum and energy within the team. This suggests that team members are motivated and engaged, which contributes to productivity and progress in the project.
2. Trait Solver-ng Development:
Team members expressed satisfaction with the progress made in Trait Solver-ng development. This signifies successful implementation and recognition of its value in the project. The efforts invested in developing and refining Trait Solver-ng were appreciated.
3. Growing Understanding of Type System Implementation:
The survey revealed a growing number of team members with a deeper understanding of the type system implementation. This is considered a positive trend as it signifies an increase in expertise and knowledge within the team. The presence of a knowledgeable team enhances collaboration and decision-making processes.
4. Stabilization, Bug Fixing, and Mir-Formality Modeling:
The team acknowledged the importance of stabilization, bug fixing, and mir-formality modeling. These aspects are crucial for ensuring a reliable and high-quality language. The survey indicated a collective commitment to address these areas and improve the overall robustness of the project.
Discussion Points:
1. Team Communication Mechanisms:
The survey highlighted several aspects of team communication mechanisms that require attention. Deep dive meetings were recognized as potentially useful but sometimes felt unproductive, possibly due to a lack of preparation or documents requiring extensive context. The single Zulip stream, combining detailed discussions about short-term goals with speculative discussions on longer-term features, was identified as challenging to follow. Additionally, there was a perception that a significant amount of communication happens on a one-on-one basis, hindering its dissemination to the broader group. However, projects like wg-trait-system-refactor were cited as examples of effective public planning, providing clarity to both team members and external stakeholders.
To improve team communication, the following ideas were suggested:
- Introduce a deep dive document template and provide clearer guidelines to enhance the productivity of deep dive meetings.
- Establish a regular knowledge-sharing meeting specifically focused on Trait Solver-ng, mir-formality, and other relevant topics.
- Consider creating separate communication streams for different topics to streamline discussions and improve accessibility.
2. Progress of Stabilizing New Features vs. Improving the New Solver:
The survey revealed differing perspectives on the balance between stabilizing new language features and improving the new solver. While some expressed concerns about potential unsoundness resulting from stabilizing new features like TAIT, others emphasized the need to avoid stalling development until the new solver is ready. The survey sought to explore options to improve this situation.
Possible approaches to address this challenge were not explicitly mentioned in the survey. However, further discussion and exploration of options, such as formality modeling, could help find a suitable balance between progress and stability.
In conclusion, the retrospective survey highlighted positive factors such as a sense of momentum and energy, progress in Trait Solver-ng development, and the growth of expertise within the team. It also identified areas for improvement, specifically team communication mechanisms and the balance between stabilizing new features and improving the new solver. By implementing suggestions such as a deep dive document template, knowledge-sharing meetings, and separate communication streams, the team can enhance collaboration and overcome challenges. Further exploration of options is needed to ensure a balanced approach to progress and stability.
Bullet points (tl;dr, too much fluff):
Yay team:
* sense of momentum and energy
* Trait Solver-ng +1
* growing set of people who understand type system impl
* stabilization/bug fixing/mir-formality modeling
Discussion points:
* team communication mechanisms
* status quo:
* we have deep dive meetings which can be useful but sometimes felt unproductive, perhaps due to lack of prep or documents that required a lot of context to understand
* single Zulip stream combines detailed discussion about short term goals with speculative discussion of longer-term features etc, can make Zulip hard to keep up with
* a lot of communication happens 1:1 and may not make its way to broader group
* *but* wg-trait-system-refactor and other projects were cited as doing a good job of working in public
* "wg-trait-system-refactor has a lot of public planning which is great :) it makes it more clear to people both on and off the team what work is involved, how long it will take, and what progress is being made."
* lang team interaction?
* more deliberate onboarding and growing of team -- have to overcome inherent difficulty of the area
* ideas for what we could do
* deep dive document template, clearer guidelines
* repeating trait-solver-ng / mir-formality / etc knowledge sharing meeting
* separate streams
* Progress the language vs new solver
* status quo:
* several people expressed concern about stabilizing new features (e.g., TAIT etc) that it would lead to unsoundness
* ideas for what we could do??
* formality etc?
* lang-team interaction
## things going well
* moving ahead on small lang inconsistencies
* TAITs and related stuff
*
* text meetings on zulip help for them being accessible
* interaction of lang + types
* is there enough, do we need other mechanisms?
* good that types can address detailed questions independently +1
* triage meetings
* but we could use more
* in-person meetups were super useful
* but we could use more
## things could be better
* new features being stabilized while trait solver is under development is stressful +1
* some meetings felt unproductive, would need more prep
* interact more with lang team
* tracking ongoing work
* chalk maintenance