Hacker News Reader

But what would sqrt(-1) output ?

Also, we are seeing the first papers trying to do numerical analysis of posit computations and... it is not better than IEEE-754, just different. Probably a better default for 16 bits arithmetic but highly debatable otherwise.

sqrt(-1) should send “You fool” to the nearest cloud printer

for the curious https://www.sigarch.org/posit-a-potential-replacement-for-ie...

Posits as defined in that article don't distinguish between +inf and -inf, so min/max operations cannot even be properly defined (since min(-+inf, 0) = {-inf, 0}).

NaNs make sense in number-crunching scenarios, but they absolutely don't make sense in anything even remotely resembling storing/communication. Worse than NULLs, really: at least those are equal to themselves.

I personally think perhaps having two kinds of float/double, one for arithmetics, with full IEEE754-compliant semantics, and another for general-purpose code, where NaNs would be just normal values that you can compare with themselves and get "true", would make sense.

JSON has no standard way of encoding NaNs. I'm never sure if this is a bug or a feature.

Which is why a lot of systems actually use https://json5.org/