You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository was archived by the owner on Feb 3, 2022. It is now read-only.
A common theme has emerged in recent threads about how the Haskell Platform is presented on the Haskell.org downloads page. It tends to go something like:
I tried using Haskell Platform a couple years ago, and it was terrible! I couldn't install the libraries I wanted to use. Why on earth does Haskell.org continue to recommend this?
For folks who are Fully Immersed in the goings-on of HP and its recent evolution, this is frustrating because we know the criticism is out of date. However, can we really fault people for being confused? To summarize:
What we now call "Haskell Platform Full" (HPF) used to just be called "Haskell Platform"
HPF/legacy HP include packages beyond the GHC-bundled packages, which introduces potentially problematic constraints when solving for new libraries
"Haskell Platform Minimal" (HPM) includes the same executables as HPF, but in terms of goals is more like an evolution of the now-defunct minimal installers
People burned by solver errors with legacy HP see the words "Haskell Platform" and assume HPM is going to give them the same problems
Let's bikeshed up a different name for HPM, so that the bad experiences people once had with legacy HP will stop hampering adoption of a product that is not at all subject to those same issues. To kick it off, a couple of ideas from recent discussions:
Core Haskell
Haskell Kit
snoyberg, queertypes, jmct, pchickey, cdsmith and 6 morellelf