[Haskell-cafe] What ever happened to Haskell 98 as a "stable branch"?

Simon Peyton-Jones simonpj at microsoft.com
Mon Mar 26 04:20:59 EDT 2007

| Why, then, are we so paranoid about introducing breaking changes in
| the development branch of haskell?  Why are we stuck without the class
| system extension proposal?  Why is Num so still so horribly mangled?
| Why can I not 'map' over a Set?  Why must I use lists of characters if
| I desire standard sorting?  Why can I not get a good error message
| from read?

It's not paranoia about breaking changes.  It's simply because no one has done it.

Haskell is rather a Darwinian sort of place.  The way lies entirely open for you to make an alternative Prelude that is just as you want it to be.  Even H98 allows you to say "import Prelude ()" to kill the Prelude, but GHC's -fno-implicit-prelude flag goes further in supporting alternative Preludes (see the docs).

OK, say you do this.  If your setup is sufficiently compelling, people will use it.  Yes, they will have to add a one-line pragma to the top of their modules, but as Alex has been arguing in another thread this weekend, it is no bad thing for a module to advertise the language in which it is written.  I don't think that one line would be a sufficient discouragement if your Prelude offered real advantages.  And Cabal makes it easy for people to download and build your library.

But building a well-engineered library, or set of libraries, is a lot of work.  That's the problem, not paranoia.

So you don't have to persuade some committee that you are right; you can just go do it.  And that would be a great way to contribute.


PS: The same goes for this suggestion

| [Conjecture 1 (2007). Haskell Mathematical Prelude and Mathematicians] If
| Haskell had a mathematically sound prelude then more mathematicians would
| use Haskell.

A mathematically sound Prelude would be great.  Go for it!

More information about the Haskell-Cafe mailing list