23 Dec
                
                    2005
                
            
            
                23 Dec
                
                '05
                
            
            
            
        
    
                3:59 p.m.
            
        Daniel Carrera writes:
If the results could be trusted, they would be useful. You could balance the expected loss in performance against other factors (e.g. speed of development).
How do you measure the time it takes to come up with a QuickSort algorithm that, implemented in Haskell, crushes the MergeSort algorithm all other languages use? ;-) Peter