Tweetovi

Blokirali ste korisnika/cu @andrasKovacs6

Jeste li sigurni da želite vidjeti te tweetove? Time nećete deblokirati korisnika/cu @andrasKovacs6

  1. 16. sij

    Did some benchmarking: Preliminary takeaway is that modern JIT platforms suck at lambda calculus.

    Poništi
  2. 11. sij

    IMO, a nice observation in this paper is that we can specify QIITs/GATs without ever touching nasty raw syntax.

    Prikaži ovu nit
    Poništi
  3. 11. sij
    Prikaži ovu nit
    Poništi
  4. 8. pro 2019.
    Poništi
  5. 24. stu 2019.

    I have an Agda file containing some large countable ordinals and a big number: Motivated by , but now my number is so large it would be a major challenge to prove statements about its size.

    Poništi
  6. 23. lis 2019.

    I previously thought that proving impredicativity consistent requires meta-level impredicativity.

    Prikaži ovu nit
    Poništi
  7. 23. lis 2019.
    Prikaži ovu nit
    Poništi
  8. 8. lis 2019.

    I have a nice and general solution for the notorious "impredicative polymorphic" type inference problem, in a dependently typed setting.

    Poništi
  9. 1. lis 2019.

    The trick is basically building binary trees from natural numbers, and after that it's pretty easy to represent other finitely branching trees.

    Prikaži ovu nit
    Poništi
  10. 1. lis 2019.

    Fun thing: it seems that all finitary inductive types are constructible from Pi, Sigma, universes, identity and natural numbers. No quotients or W-types needed. UIP/funext perhaps needed for some types but not for others.

    Prikaži ovu nit
    Poništi
  11. 15. ruj 2019.

    The answer which I've found is: if we add extensional equality to Jon Sterling's cumulative algebraic TT, internal cumulative subtyping is derivable. This seems to be the nicest formalization.

    Prikaži ovu nit
    Poništi
  12. 15. ruj 2019.

    This can be used (tedious, didn't formalize, but obviously works) to model Jon Sterling's cumulative algebraic TT.

    Prikaži ovu nit
    Poništi
  13. 15. ruj 2019.

    Transfinite cumulative Russell-style inductive-recursive universes in Agda:

    Prikaži ovu nit
    Poništi
  14. 5. ruj 2019.

    "Propositional subtyping" would behave the same as extensional prop. equality. I.e., from a term proving propositional subtying, I get judgmental subtyping by "subtyping reflection".

    Prikaži ovu nit
    Poništi
  15. 5. ruj 2019.

    Question to twitter. Cumulative universes induce subtyping in Coq: I'm using this kind of cumulativity informally, but want to internally *prove* subtyping between some complicated types. Is there prior art for this "propositional" subtyping?

    Prikaži ovu nit
    Poništi
  16. proslijedio/la je Tweet
    31. kol 2019.

    This will be fun. Anyone want to come and hack on Idris with me?

    Poništi
  17. 15. lip 2019.

    I'm at ZuriHac now, is there anyone who wants to chat?

    Poništi
  18. 27. svi 2019.

    He's going to give an invited talk about generalized sketches next week at our type theory club, I'm pretty excited about it.

    Prikaži ovu nit
    Poništi
  19. 27. svi 2019.

    We belatedly found out last week at our university, that in fact Makkai has been teaching categorical logic for the whole past semester at the philosophy institute (!).

    Prikaži ovu nit
    Poništi
  20. 13. svi 2019.

    Our shallow embedding is generally very helpful, but for purely syntactic translations it's absolutely killin it. I'm excited about what we could formalize, maybe I'll revisit my stuff on closure conversion for dependent types.

    Prikaži ovu nit
    Poništi

Čini se da učitavanje traje već neko vrijeme.

Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.

    Možda bi vam se svidjelo i ovo:

    ·