I believe Simon got right what he got right, but nothing more. Haskell is an experimental language that is taking off prematurely. In my opinion it needs to be taken back to the drawing boards and overhauled before it is too late. It may already be too late unfortunately. We may have to live with whatever mistakes that were made for the next 100 years. Sometimes success is not a good thing. Haskell can easily become yet another defective language. Backward compatibility will likely need to be thrown out the door if Haskell is to ever stand for the "truth". Greater rigor in getting it right is needed.
Haskell will die with the two Simons. The future lies with Agda or Epigram or Coq. We shouldn't try too hard to get Haskell right -- it will never be right -- but, rather, we should all think about how we're going to migrate away from Haskell to something even better. What can we do now -- in terms of multi-versioning, compiled module interfaces and approaches to modularity that are about servers and pipes instead of linking -- to make it easy to move away from Haskell module by module? -- Jason Dusek
participants (2)
-
Jason Dusek -
John D. Earle