I’ll try once again (this is the last one, I’ll promise) to convince the Committee of including ScopedTypeVariables as part of GHC2021. It has been with us for a long time — so their problems seem well understood — and the community seems to want it as part of this new “language version” — which means that people are really using it.

Most of the criticisms are around the fact the the current manifestation is not ideal, something I agree with. I would say in this case we should not let perfect be the enemy of good: if during the years, once types in patterns or Dependent Haskell extensions require this, the scoping mechanism needs to change, we can do it. But as of now, ScopedTypeVariables is a very important piece in the flavour of Haskell in main usage (given its mainstream acceptance).