Skip to content

Conversation

@pedagand
Copy link
Collaborator

I remove the Set\omega here and there, and then deploy fold (modern version) for GNDT.

First, TT is not actually using the fact that is defined in Setω. It
could just as well have been defined in, say, Set₀ and all examples
would still work.

Second, the definition of Map (and subsequent Fold, etc.) were highly
suspicious: they were significant size changes between sources and
targets. I'm not sure these are still related to the actual
categorical concepts.

This also manifests itself in the type of induction for LNDT, which is
uselessly generalized wrt. to the Set level of `A`.

Overall, having a down-to-earth definition for TT (and elsewhere) will
make my life easier.

(And premature universe level generalization is the root of all evil)
We can go beyond list-like containers now. This slight generalization
supports any kind of sum-of-product, first-order datatype. This could
be further extended to support Pi-types and Sigma-types, à la Desc
described in

  https://www.irif.fr/~dagand/stuffs/thesis-2011-phd/thesis.pdf
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant