ajt
(Alex Townsend)
1
If I type this into wolframAlpha
sum log(k)/(k) for k = 2..n
then I get
StieltjesGamma(1)-?_1(1+n)
where it also tells me that ?_1 is the StieltjesGamma(1) function. (It’s strange that it did not simplify, but OK.)
However, when I go to verify the equation at n=100, it does not work out.
StieltjesGamma(1)-StieltjesGamma(1)*(1+100)
7.2815845483676724860586375874901319137736338334337952
but
sum log(k)/k for k = 2..100
10.553976183549153710596114240658388851398614872416869
Is this a bug? It does not make much sense to me. Any ideas what is going on?
mthiel
(Marco Thiel)
2
Hi,
Sum[Log[k]/k, {k, 2, N}]
(*StieltjesGamma[1] - StieltjesGamma[1, 1 + N]*)
Note that the second time the functions called with two arguments. If you evaluate:
StieltjesGamma[1] - StieltjesGamma[1, 101] // N
(*10.554*)
which is the same as
Sum[Log[k]/k, {k, 2, 100}]
You are typing in
StieltjesGamma[1] - StieltjesGamma[1]*(1 + 100) // N
instead with evaluates to 7.28158.
So this is not a bug. You just evaluated different function. Note that StieltjesGamma is overloaded: you can use it with one or two input slots.
Cheers,
Marco