Re: question on a possible new function or extending the existing one for core()

[Aurel Page <aurel.page@normalesup.org>]

Dear Randall,

Since factor accepts t_FRAC inputs, we could consider extending core, 
but it is not clear to me that the definition you propose is meaningful. 
qcore2(n) = core(numerator(n))*core(denominator(n)) would give unique 
representatives in Q*/(Q*)^2, unlike your definition. What is your 
motivation for this extension?

Cheers,
Aurel

On 21/05/2025 10:36, American Citizen wrote:
> We know that core finds the non-square portion of an integer number 
> and returns it, or if a nonzero flag is used, returns both the 
> non-square and the positive square root of the square (which is handy)
>
> Can we extend core to the t_FRAC type? I defined qcore(n) = 
> core(numerator(n))/core(denominator(n)), and I suppose f can be found 
> from this too.
>
> Please consider extending this function? Thank you.
>
> Randall