Burning bridges

[wren ng thornton]

On 5/20/13 6:59 PM, Felipe Almeida Lessa wrote:
> IMHO, the problem is that the community isn't large enough to be able
> to suffer a split.

That is indeed the problem. Which is why we've argued about it forever 
without making any progress, IMO.


> If we're going to burn bridges, perhaps we should follow a
> Python-3-esque path of releasing a major upgrade together with:
>
>    - A refactoring tool to aid with the transition.
>
>    - A deadline of, say, 2-3 years during which the latest GHC for
> current Haskell would receive bugfixes while GHC 8 moves along as
> usual.

That's actually what I had in mind, rather than a true (parallel) fork. 
Just an official jump from Haskell-1 to Haskell-2, with no claims of 
compatibility between the two.

Of course, if we are going to do this, I think it'd be good to make the 
decision official and then take a year or so to work out the full 
details. That is, if we're going to break everything, we should make 
sure we don't overlook any of the low-hanging fruit[1]. Given the recent 
turnover for the haskell' committee, it'd be good to talk to them too. 
Even with a big jump, I think it'd be good to aim for conservatism; that 
is, the goal is to remove all the major warts, not to solve any deep 
problems.


[1] One that has been bothering me recently is the requirement that 
methods of Enum are partial functions. I have a proposal for fixing this 
and generalizing enumeration, which I'll post as a package later this week.

-- 
Live well,
~wren