News for hackers. Stuff that matters.

A comment on 1/0 = 0. Parent comment

Re:1/0 = 0 by jibal on Saturday August 11, 2018 @04:17AM (#17738873)
"In this case, it's "obvious" that when e.g. a is zero, you want 1/a to be zero so you simply return b"

b obviously isn't what you want from this harmonic mean when a is 0. And in any case x will be 2b if 1/a is 0, not b.

[ Parent | on HN ]
Re:1/0 = 0 by semi-extrinsic on Saturday August 11, 2018 @10:03AM (#17739709)
So, I did forget the correction such that you don't get 2b. The general case is described e.g. in the link below, and this is known as a "zero-corrected harmonic mean". It is actually "obvious" in some cases, but I'll concede that in other cases it's not.

https://www.rdocumentation.org/packages/lmomco/versions/2.3....

[ Parent | on HN ]
Re:1/0 = 0 by jibal on Sunday August 12, 2018 @03:02AM (#17743299)
Eh? That page says nothing about 1/0 being 0, or anything like that. That function takes the harmonic mean of the non-zero samples, and then adjusts the result by multiplying it by the fraction of the samples that are non-zero. It purposefully avoids dividing by 0. And the result of that function for a==0 and b!=0 is b/2, not b or 2b. So not only is what you claim to be obvious not obvious, it's flat-out wrong.
[ Parent | on HN ]