Did not realize that 0 * Infinity is NaN. Anyone know why that wouldn't be expected to be 0?
-
-
Replying to @sgrif
0 * x = 0, Infinity * x = Infinity -- who wins? NaN it.
2 replies 0 retweets 0 likes -
Replying to @Gankra_
x already has additional bounds in the infinity case though. Could just as easily be Infinity * x = Infinity when finite, nonzero
1 reply 0 retweets 0 likes -
Replying to @sgrif
This is the general pattern of indeterminate forms; two "rules" apply, so we throw up our hands in disgust: https://en.wikipedia.org/wiki/NaN#Operations_generating_NaN …
2 replies 0 retweets 0 likes -
Replying to @Gankra_
Fair enough. I guess I assumed that the math backed up the 0 case, but it appears not to
1 reply 0 retweets 0 likes -
Replying to @sgrif
the fact that Infinite is a "value" and that 1/0 produces it is already a fairly disgusting abuse, depending on who you ask.
3 replies 0 retweets 0 likes -
Somewhat more defensible in the context of IEEE754 if you think of 0 as also including underflow and Inf including overflow.
2 replies 0 retweets 0 likes -
my general rule of thumb on these things is "do standard algos degenerate correctly with this behaviour". No clue here.
2 replies 0 retweets 0 likes -
IEEE 754 is very much a pile of "OK most of the time" compromises.
1 reply 0 retweets 0 likes
That's more or less computers in a nutshell really. :)
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.