GSoC and RustReach Day 13

May Day Convoluted

• Phil gave a talk similar to his DOM talk, but he interestingly mentioned a few papers, including Day’s Convolution and Arthur Xavier’s Comonads/UI thesis
• Cofree meets Comonads
• The dual of substitution is redecoration - Uustala and Vene

Notes

• (Nate’s talk)
• I actually understood a lot of it! :D It was more beginner-friendly, and on reusable code.
• Reusable Code
• Generalization of functions
• type v class..typeclass
• using => compiler takes care of implementation
• Apply makes functor composable for any implementation of Functor
• Final encoding (provide implementation as argument)
• Initial encoding (interpret a result)
• Aff is opaque: you have to run it to get what it contains (ie not reusable)
• How can we make it reusable?
• Datatype that represents semantics of Aff

Data RealCode
= ReadFile String
 | WriteFile String String
 | HttpGet String
 
vs

...ReadFile String (Maybe String -> a)
...WriteFile String String (Unit -> a)
...HttpGet String (String -> a)
| Done a

• ~> is a natural transformation
• extensible datatypes let you combine types without boiler plate (lends themselves to more reusable code)

Rust

• I met with my mentor today. He was in Paris at the RustConf there, so our internet was a bit spotty, but he really helped me with understanding Traits. My mission this week is to continue from chapter 11
• traits: a way to share methods between two structs
• passing traits : IO
• Hash keys
• Key has to implement Hash
• I’ve been told that by next week, it’s sort of Discovery, where I see what issues I’d like to fix. Neat!

Back to May Day Convoluted

• Chris B.
• ontology: how do we define identity? Using Category Theory for ontological relationships

Andrew

• Monoids on Steroids
• 1942: Category theory invented
• 1958: Lisp invented
• 1983: Symbolics
• Lisp machines: latent typing, tagged architecture
• XML: typed S expressions (closing tag has to be enforced)
• Bill Gospher
• Julia sets