semantic/weekly/2016-05-24.md

May 24th, 2016

NB: On Tuesday this week since Monday was Victoria Day.

Agenda

1. @robrix to describe a couple of tweaks he’d like to make to this meeting, followed by discussion of same:
   • Adding “what did you enjoy” question.
   • Rotate facilitator & note-taking tasks weekly.
   • Going through the retro questions point by point instead of person by person.
2. Retrospective on last week:
   • What went well?
   • What was challenging?
   • What did you learn?
   • What did you enjoy? (If we didn’t decide not to do this.)
3. (Meta)retrospective on the format &c.

Tweaks to the format

• 👍 to trying it out.

What went well?

@joshvera:

• Diff summary meeting on Wed.
• Finished up a lot of the skeleton of diff summaries.
• Yesterday paired w/ @rewinfrey & shared context about summaries.

@rewinfrey:

• Pairing w/ @robrix last week.
• Asking questions!
• Learned a lot.
• Came out of the week feeling hopeful rather than defeated.
• Pairing w/ @joshvera yesterday on diff summaries. Got context about shapes, helped solidify intuitions re: the datatypes &c.

@robrix:

• Pairing!
• Made a lot of progress on the alignment stuff, which was a secondary goal, so it’s a pretty great bonus.
• Particularly interesting building the infrastructure to do property tests around alignment stuff.

What was challenging?

@joshvera:

• Stubbing out parts of diff summaries that we don’t have good answers for. E.g. productions, what type of info we’ll get out of parent annotations/contexts when constructing summaries.

@rewinfrey:

• Newness to Haskell & the project. Trying to solidify understanding of Haskell at the same time as the shapes of the data, especially in contexts like the Alignment problem. Challenge is diminishing over time.

@robrix:

• Pairing as much as we did was exhausting, just because I have to be “on” for extended periods of time; worried that I tired @rewinfrey out too. Feel like this will be ameliorated organically what with long weekends and schedule tweaks and @joshvera being freed up a bit. Also just getting used to it will help.
• There are some question marks about parts of the alignment stuff that are making me a bit nervous. Adding test cases for a couple of those will definitely help!

What did you learn?

@joshvera:

• There’s a ton more stuff to do to get the types of diff summaries that we actually want as opposed to the ones that our current system produces, e.g. w.r.t. tree-sitter.

@rewinfrey:

• Learned a lot.
• The transition from thinking about function application to thinking about function composition was an eye-opener.
• Understanding a lot more about why certain shapes are advantageous.

@robrix:

• The halting problem applies to human analysis of programs too; sometimes you still have to go through a problem step by step in your head to understand it.
• Analyzing infinite loops in terms of general recursion/primitive recursion is pretty clarifying; e.g. “why isn’t this primitive recursive? Can this be strictly positive?”
• Reminded that I don’t really understand a thing til I’ve articulated it clearly.

What did you enjoy?

@joshvera:

• Pairing w/ @rewinfrey yesterday.
• Giving the demo Wednesday.
• Voicing how I think about the structures we use.

@rewinfrey:

• I feel like everything is pretty awesome.

• Getting paid to do Haskell is a dream come true.

• Haskell honeymoon.

• Learning.
• Working w/ @joshvera & @robrix. Patience; willingness to share; egolessness. Ed: 💟

@robrix:

• Pairing w/ @rewinfrey.
• @rewinfrey’s questions!
• Articulating approaches taken; e.g. the meta-level of the function application to function composition transition. I hadn’t really been conscious of that transition until @rewinfrey asked why I used pointfree style in one part of the code, or how I decide whether to write a function tacitly or not. Being forced to articulate it helped me understand that this had happened, which in turn sheds light on the difference between “can” and “should” for this particular problem.

Metaretro

• @joshvera: Some of the questions felt like I was repeating thoughts by rephrasing through a different lens.
• @rewinfrey: I think I might keep these questions in mind throughout the week.
• @robrix: All of the above.

In response, going to drop the “What did you enjoy?” question, & have the facilitator/note-taker for the next meeting open the agenda PR for it the week before (i.e. today).