HP Prime Infinite Sum Bug?

[kcnicho]

Hi All,

I posted this question over at the HP Calculator site (http://h30499.www3.hp.com/t5/Calculators...-p/6662418), thought I'd ask here as well.

HP Prime, in CAS mode, when I enter "sum(1/sqrt(n), n, 1, infinity)", the Prime returns a screen full of warnings (see attachments.)

I believe the correct answer should be "+ infinity", as I get from Wolfram Alpha as well as the TI-Nspire CAS.

Has anyone else seen this? Can anyone else duplicate this behavior on their own machine? I verified this on the Prime Virtual Calculator, Version 2014 3 31, Rev 6031 (latest firmware, as far as I know.)


Thanks,

- Kevin

[parisse]

Sorry, convergence of series is not yet implemented in giac. I plan to do that as part of the new step by step feature.
sum tries to find a closed form for the discrete antiderivative (that's why you see all these messages), it fails, hence returns the sum unevaluated.

[Adriweb]

(11-05-2014 12:34 PM)parisse Wrote:  part of the new step by step feature.

Oh, that's exciting
(Coming to KhiCAS too ? )

[parisse]

If I can find a way to display intermediate computations on the nspire screen.

[kcnicho]

Thanks for the (definitive) answer!

Happy to hear that things are in the works to address these types of problems in future firmware releases.

- Kevin

[salvomic]

(11-05-2014 12:34 PM)parisse Wrote:  Sorry, convergence of series is not yet implemented in giac. I plan to do that as part of the new step by step feature.
sum tries to find a closed form for the discrete antiderivative (that's why you see all these messages), it fails, hence returns the sum unevaluated.

I hope you could do it soon in a next firmware upgrade, thank you!
It would be very interesting...

Salvo

[salvomic]

...another Series I would se calculated in Prime:
sum((-1)^(k+1)/K^2, k, 1, infinity)
It should give π^2/12

Thank you
Salvo