Re: [Help-gsl] looking for gsl_integration_long_double

[Evgeny Kurbatov]

> > Please tell, what the task you solve with such requirement to precision?
>
> Well, does it matter what reason Tobias has? He is looking for a 'long
> double' integration and there isn't any. Sounds like a good reason to
> me.
>
> *My* reason for wanting such a thing would be simply that all the rest
> of the program is using long double in the numerics part and sometimes
> the easiest way to check for error propagation is to require higher
> precision and check for the changes in the result.

No, I just have an academic interest in features of such tasks and I 
didn't try to reject the request of Tobias.

BTW I believe that in numerical modelling the algorithm must not work in 
a regime of the limit of prescision.

Regards,
Evgeny