The dreaded M-R

[John Hughes]

Philippa Cowderoy wrote:

>On Mon, 30 Jan 2006, John Hughes wrote:
>
>  
>
>>Wouldn't there be quite a lot of warnings?
>>
>>I really think we have to remember beginners' problems here. I know noone
>>reading this discussion is
>>one, but without beginners, there will be no growth in the Haskell community.
>>If we did as you suggest,
>>then beginners would be confronted by yet another kind of incomprehensible
>>message--and I fear it
>>would be hard to report that x=e risks evaluating e many times, in a way that
>>would be viewed as positive
>>by someone still trying to understand how Haskell works at all.
>>
>>    
>>
>
>Insist the warnings be available, don't require them to be in the standard 
>warning level?
>
>  
>
But then beginners will, without warning, sometimes find their code 
running unreasonably
slowly. That isn't, really, any better. Too many will just conclude 
"Haskell is unreasonably
slow" and abandon it.

John