[Haskell-cafe] Monad of no `return` Proposal (MRP): Moving `return` out of `Monad`

[John Wiegley]

>>>>> Henrik Nilsson <Henrik.Nilsson at nottingham.ac.uk> writes:

> Of course 25 years+ of legacy is there. Quantifiable or not. And of course
> it must count if we wish Haskell to be taken seriously and be increasingly
> successful. That is at least the view of a significant part of the Haskell
> community. And the way to account for it, both that which is quantifiable
> and that which is not, is to make sure that fundamental language changes are
> given due consideration in a forum where all stakeholders are adequately
> represented.

Henrik,

I do appreciate this point of view. It must be stated, however, that some have
become frustrated by the frozen state the Haskell language has entered. No
substantial changes have occurred since 1998. A new Prime committee is being
nominated, it is true, but that has happened before.

Arguments for stability might be seen by some as arguments to never change,
since the same pain we must endure now, will also have to be endured later.
Books and code will still be broken, teaching will still be made harder, etc.
A willingness to accept that pain shows concretely that a brighter future is
coming.

I'm very much against frivolous change, but I strongly want improvement. If it
means a progressive delivery of changes, or grouping them a bit more, that is
fine. I just worry when I see the "caution" argument used to put a break on
daring changes time and again, because otherwise, why should we change at all?

By comparison, the C++ language -- significantly more complex and affecting a
much larger community of users -- has released 4 separate standards in the
timeframe between Haskell 98 and now, each of them introducing significant
functionality that created work for compiler and tools vendors, users,
authors, teachers, and learners. If they can do it, I'm confident we can as
well.

John