[HN Gopher] Type Inference Zoo
       ___________________________________________________________________
        
       Type Inference Zoo
        
       Author : mpweiher
       Score  : 153 points
       Date   : 2025-06-18 08:02 UTC (4 days ago)
        
 (HTM) web link (zoo.cuichen.cc)
 (TXT) w3m dump (zoo.cuichen.cc)
        
       | flashgordon wrote:
       | This is amazing. I've been wanting to do this for ages (even
       | tried a few times and failed a few times). Great to see this.
        
       | dmkolobov wrote:
       | It is crazy how much work comes out of Bruno Oliveira's research
       | groups.
       | 
       | Another cool online thing that exists is CP, a language with
       | disjoint intersection and union types:
       | 
       | https://github.com/yzyzsun/CP-next
        
         | gavinray wrote:
         | TypeScript has this functionality.
         | 
         | Additionally, Scala 3 does as well, including support for these
         | generic "Type-Lambda" expressions I saw in one of the examples
         | on the playground:
         | 
         | https://beta.plground.org/jnclma/CP_tutorial
         | printNum = trait implements ExpSig<Eval => Print> {
         | 
         | In Scala 3, this would be roughly:                 trait
         | ExpSig[F[_]]       object printNum extends ExpSig[[A] =>>
         | Eval[A] => Print]
         | 
         | https://docs.scala-lang.org/scala3/reference/new-types/type-...
        
       | b0a04gl wrote:
       | huh funny how python's doing the job of modeling type systems
       | when it barely enforces types itself. but makes sense the more u
       | poke around. didn't have to fight any compiler layer or deal with
       | noise. could just build the inference logic raw, decide how types
       | get unified, how ctx gets stored, all that. no annotations or
       | weird conversions. the language just stepped aside so i could
       | focus. actually helped my brain map the theory better ngl
        
         | b00ty4breakfast wrote:
         | >uh funny how python's doing the job of modeling type systems
         | when it barely enforces types itself
         | 
         | The magic of computers. You can write a standards-compliant
         | Lisp interpreter in python, too and it doesn't have a large
         | chunk of that stuff in itself either.
        
       ___________________________________________________________________
       (page generated 2025-06-22 23:01 UTC)