On Fri, 19 May 2006, skaller wrote:

> On Thu, 2006-05-18 at 14:53 -0400, Jacques Carette wrote:
>> skaller wrote:
>>
>>> What about high level optimisations?
>>>
>>> Felix supports this:
>>>
>>> 	reduce revrev[t] (x:list[t]): rev (rev x) => x;
>
>> Haskell (GHC to be precise) allows that too. But is syntactic
>> term-rewriting, in other words it is *untyped*.
>
> It's well typed. x:list[t] means x is of type list[t].

I'm just wondering how often such optimizations are worthwhile.  From what 
I've read, basic register allocation and peephole optimizations gain you 
2x the performance- after that, it's a fight for percentage points.

Especially considering that humans tend to be a lot better at recognizing 
opportunities for high-level optimizations than computers are.  For 
example, a human would be much more likely to notice a double list 
reversal when the two reversals are in seperate modules, or otherwise 
obscured from the compiler.  Even more so, the programmer may decide that 
maybe a list isn't the right datastructure for this data- a decision I 
wouldn't trust to a compiler.

Brian