Compare commits
2 Commits
7623787b1c
...
068d0218b0
Author | SHA1 | Date | |
---|---|---|---|
068d0218b0 | |||
3e9f6a14f2 |
|
@ -361,7 +361,7 @@ I think this is a good approach. Should we want to add more types to our languag
|
|||
lists, and so on, we will be able to extend our `decEq` approach to construct more complex equality
|
||||
proofs, and keep the `typecheck` method the same. Had we not used this approach,
|
||||
and instead decided to pattern match on types inside of `typecheck`, we would've quickly
|
||||
found that this only works for types with finitely many values. When we add polymorphic tuples
|
||||
found that this only works for languages with finitely many types. When we add polymorphic tuples
|
||||
and lists, we start being able to construct an arbitrary number of types: `[a]`. `[[a]]`, and
|
||||
so on. Then, we cease to be able to enumerate all possible pairs of types, and require a recursive
|
||||
solution. I think that this leads us back to `decEq`.
|
||||
|
|
|
@ -54,6 +54,10 @@ div.highlight table {
|
|||
code {
|
||||
border: none;
|
||||
}
|
||||
|
||||
span {
|
||||
margin: 0 !important;
|
||||
}
|
||||
}
|
||||
|
||||
.container {
|
||||
|
|
Loading…
Reference in New Issue
Block a user