Growing Haskell Platform

Vincent Hanquez tab at snarc.org
Fri Dec 7 15:11:53 CET 2012


On Thu, Dec 06, 2012 at 11:24:41PM -0800, Mark Lentczner wrote:
> In my Haskell Implementors 2012 talk Haskell Platform: Field Report and
> Future Goals, I called for an expansion of the content of the platform to
> bring it up to par with what other programming language platforms provide.
> 
> Please see my blog post today:
> http://mtnviewmark.wordpress.com/2012/12/06/growing/)
> 
> Where I outline what is missing, and call for a ramp up in our efforts to
> grow the platform.

Hi Mark,

I agree in general with the goals of growing the haskell platform; I more than
once expressed the same interests.

However I really don't see how it can happens at the moment.

The bar is too high in general. While i can understand the reason (and agree)
for most of the requirements on the list of acceptance, in general this is
working against the platform. I would prefer the platform to offer many more
good packages that maybe doesn't meet all the high quality requirements, but
provide a necessary feature for users, and see them improve over time through
the platform.

At the moment, I think we expect perfect packages that will enter a freezing
state after being accepted.

In my mind, having multiple levels of platform could be a good idea. the
current platform would be platform-base, and you could have something like
platform-unstable, platform-experimental, that would mark the intent of being
in the platform:

- users can start rely on the package with the clear expectation of change in
  next versions.
- users/devs can proactively improve packages that are missing stuff

I think there's clearly a middle ground between the graveyard-like property of
the platform and pure cabal/hackage usage. I believe this is being worked on,
from the other end with Michael's stackage, but more effort to meet in the
middle would be a good idea.

-- 
Vincent



More information about the Libraries mailing list