You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Many lemmas in HITs.SetQuotients.Properties, notably elimProp[n], elimContr[n], and setQuot[n]Op, could be generalised so that:
The quotient relation R doesn't have to have the same type as A
When n > 1, the type family can depend on different quotients (e.g. C : A / R → B / S → Type ℓC
This is tedious work, but not particularly difficult, and it's useful. For example, when I was defining quotient categories, I couldn't use these lemmas, since they required the quotient types to all be the same. Hence I had to define operations/axioms on the quotient from scratch (using elimProp), which could have been avoided with more general versions.
I should also mention that the proofs are likely to be identical: only the type would be changed.
The text was updated successfully, but these errors were encountered:
Many lemmas in
HITs.SetQuotients.Properties
, notablyelimProp[n]
,elimContr[n]
, andsetQuot[n]Op
, could be generalised so that:R
doesn't have to have the same type asA
n > 1
, the type family can depend on different quotients (e.g.C : A / R → B / S → Type ℓC
This is tedious work, but not particularly difficult, and it's useful. For example, when I was defining quotient categories, I couldn't use these lemmas, since they required the quotient types to all be the same. Hence I had to define operations/axioms on the quotient from scratch (using
elimProp
), which could have been avoided with more general versions.I should also mention that the proofs are likely to be identical: only the type would be changed.
The text was updated successfully, but these errors were encountered: