@defnotbeka Type holes, first class versioning, built-in API expand-contract, ability to deprecate/inline code into client code-bases, report usage statistics of APIs, facilities for state migration, no global namespaces, some form of differential programming for custom automatic patches
I’m looking for participants for a paid user study on a prototype system for tabular programming (programming with tables / data frames). 2 hours • $50 Visa prepaid card Hoping to finish sessions by March 13. Details + sign up in 🧵 Reposts appreciated! 🙏
Your participation will involve completing programming tasks using the prototype. We’re especially looking for folks with experience in statically typed functional languages (OCaml, Haskell, Scala, F#, Rust, etc.).
Sign up: https://t.co/n8XuvsUcTe
I’m looking for participants for a paid user study on a prototype system for tabular programming (programming with tables / data frames). 2 hours • $50 Visa prepaid card Hoping to finish sessions by March 13. Details + sign up in 🧵 Reposts appreciated! 🙏
an interesting update: the team is starting to move away from AI coding completely (devin/claude/etc) because it's so much harder to review the AI code than writing things themselves
@Keleesssss@HSVSphere@steinerkelvin@_hakagie@disconcision So in order to type `if` as a variable you need to complete the surrounding context and prefill any whitespace then go back and insert the text and navigate away without typing so it doesn't get remolded. And if anyone inserts whitespace there it re-expands to an if expression.
@Keleesssss@HSVSphere@steinerkelvin@_hakagie@disconcision So there is a unmolded node that gets turned into a partial AST but this happens from inferring context based on the inserted characters. So if you type `if` and then write a space to the right of it it gets molded into a if expression.
@Keleesssss@HSVSphere@steinerkelvin@_hakagie@disconcision Yeah, I just think the delineation between parsing and structured editing here is somewhat difficult since everything is inferred from text. You could imagine the same situation with an incremental reconstructing parser that operated on text patches.
@Keleesssss@HSVSphere@steinerkelvin@_hakagie@disconcision Even if you’re not you’re interpreting it as a sequence of text post hoc, you’re interpreting it using a sequence of character updates that get manifested as structural changes. So you have an AST that is argue many users could not figure out how to insert.
I am once more seeking participants for a user study. We'll be evaluating a prototype code editor that completes your text with placeholders for missing tokens expected by the language. Here are a couple vids motivating the editor+study.
Details in thread. RTs appreciated! 🙏