Re: [math] NaN vs. Exception Revisited (was: Re: [math] proposed ordering for task list, scope of initial release)

[Tim O'Brien <to...@discursive.com>]

> I think your usage of Convergence Exception is appropriate in your case. 
> In your case your waiting for the method to return, if it fails to 
> converge, you need to know why, there may be different reasons, and its 
> an "exceptional case" which probibly terminates the whole process in the 
> long run.

+1, a ConvergenceException makes sense.

> try{
> 	MathUtils.factorial( Math.floor( Math.log( 	
> 		StatUtils.geomean(someobj.getArray()) )))
> 
> }catch(InvalidInputValueException iive){
> 	iive.printStackTrace(...);
> }

-1, not only is that a very bad idea in terms of usage constraints, it
tends to buck a well established trend.   Complex "algorithms" should
through checked exception, something like finding an Average of a set of
values should return NaN.



---------------------------------------------------------------------
To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: commons-dev-help@jakarta.apache.org

Re: [math] NaN vs. Exception Revisited

["Mark R. Diggory" <md...@latte.harvard.edu>]

Tim O'Brien wrote:

>>I think your usage of Convergence Exception is appropriate in your case. 
>>In your case your waiting for the method to return, if it fails to 
>>converge, you need to know why, there may be different reasons, and its 
>>an "exceptional case" which probibly terminates the whole process in the 
>>long run.
>>    
>>
>
>+1, a ConvergenceException makes sense.
>
>  
>
>>try{
>>	MathUtils.factorial( Math.floor( Math.log( 	
>>		StatUtils.geomean(someobj.getArray()) )))
>>
>>}catch(InvalidInputValueException iive){
>>	iive.printStackTrace(...);
>>}
>>    
>>
>
>-1, not only is that a very bad idea in terms of usage constraints, it
>tends to buck a well established trend.   Complex "algorithms" should
>through checked exception, something like finding an Average of a set of
>values should return NaN.
>
>  
>
Just testing the conceptual waters, if we still feel as strongly about 
it, very well...



---------------------------------------------------------------------
To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: commons-dev-help@jakarta.apache.org