[Haskell-cafe] Re: Non-technical Haskell question

[Robert Dockins]

Henning Thielemann wrote:

> On Mon, 6 Dec 2004, Robert Dockins wrote:
> 
> 
>>The problem, of course, is that Haskell likes to tightly bind with the 
>>libraries it uses (inlineing across modules and other optimizations). 
>>So imaging if the "package" unit was a barrier to those kinds of 
>>optimizations.  Then, no knowledge of the internals of the package are 
>>needed by importing modules, and "sufficently" compatable pacakges could 
>>be drop in replacements, .so or .dll style.
> 
> 
> This would mean that functions like 'map' and 'foldr' couldn't be unrolled
> because they are in the package of the standard functions?

Probably it would make sense to have a basic set of modules which are 
not in any package, but which are avaliable to import into all modules. 
  I image this would include things like the Prelude, the Haskell98 
basic modules, the new Data.* modules, Control.*, Foreign.*, and 
probably a few other things as well (in short, most of the current 
Hierarchical Libraries).  Obviously, barring map and foldr from inlining 
would be a Bad Idea.

Things which exist now that I would think would work well behind a "no 
inline" barrier:

Parsec
wxHaskell
HAXML
HToolkit
WASH
dozens of others I don't know about


I would guess that most of the entry points for these libraries are 
already too complicated for GHC to think that it is "worth" inlining 
them.  Mandating a "no inline across packages" rule would likely not 
change their performance much.