Meeting notes (3/6/2020)

Minutes

MaybeTextCodec
- Reads in 8k buffer of bytes
- If UTF-8, emits Primitive::Strings, other Primitive::Bytes
- Loading all data into memory?
  - serde and decoders to see how they play into streaming data
  - Chars iterator instead?
- collect_string and collect_binary may need some cleaning up
open /dev/urandom hangs 😬
- Async http is a thing now, could work for fetch
- Config to avoid sniffing the stream, just always have raw bytes?
- Convention wise, how can we make it intuitive
  - Windows still likes UTF-16, so keep in that mind
named pipes (load data from side channel)
- Performance concerns, named pipes could deal with this
- Works on windows too!
- Focus on things that help adoption
Lazy field resolution
- ps having port information, but only computing if needed
- Bidirectional? Pipeline analysis phase?
Exploring nu
- Exploring like in an R shell
- Data frames
  - Imagine a variable that you can ask
    - What columns do you have?
    - How many rows?
- Commands in namespaces
nushell as a data processing shell
- Parallelization? Distributed computing?
- Cron jobs?
How do we prioritize?
1. Fundamentals
  - Aliases / Functions / Variables
  - Make scripting work
  - Completions
    - ps | where <TAB>
    - git branches, for example
2. After the above, anything that helps with adoption
What is the core team?
- Diversity
- How do we bring on new people?
- How big should the core team be?
- Should there be a core team, and a committer's team?
  - Rails model

Provide feedback

Saved searches