[seqfan] sigma(n) / d(n)

[Charles Greathouse]

I added comments from Sutantyo and Bateman et. al. to A057020.  Would
someone check that these are actually consistent with each other?  The
formulas are different:
sum(k=1, n, sigma(n)/d(n) * 1/n)
vs.
sum(k=1, n, sigma(n)/d(n) * 1/k)
but the constants given do not seem to be compatible.  (Or perhaps I
am misreading them?)

Charles Greathouse
Analyst/Programmer
Case Western Reserve University