OK, well, I felt I didn't completely understand almost any of them.
So I thought the issue tracker would be a good place to clarify them.

That's quite fair. I'll try to be better.... I admit that the TO DO is my dumping ground for half-baked thoughts when I'm in the thick of coding.
 
More importantly - I'd like to know what the most important bottlenecks
are for getting the next HP out the door, and what are some
good places to jump in to help with that.

Bottlenecks in the new build:
  • The new build needs to be tested on linux. The build has a generic linux build that produces a system image under build/target but then I don't know how (or if) particular distributions would use that to build an os specific distribution package. You can follow the code in htptool/src/OS for how I did this for Mac OS X.
  • Mac OS X build needs a pile of scripts and static files added - this is so mostly, after the installer is run, the target system has symlinks in all the right places, and the user sees the welcome .html file. I'm in the thick of this one (but it is tedious and slow work as testing it is a huge pain involving multiple machines!)
  • The website needs to be templatized and built from hptool - template generate web site · Issue #92 · haskell/haskell-platform
  • The master haddock generation needs some hard love: I don't think that the internal links are all correct -- though perhaps this isn't crucial. In particular, I worry that the links from platform package to ghc packages are wrong.
  • The package list and versions need general agreement
everything else is minor and doesn't need to happen to get a release out.

Something that isn't a bottleneck, but would be a HUGH boon to HP development would be continuous build · Issue #94 · haskell/haskell-platform

- Mark