4 ms·
"Perfect Random [sic] Floating-Point Numbers" ??? I had hoped that, somewhere in the article, the author would say, "In this article, the term 'random' is shor
by lutusp 1y ago
"Perfect Random [sic] Floating-Point Numbers" ???
I had hoped that, somewhere in the article, the author would say, "In this article, the term 'random' is shorthand for 'pseudorandom'." But no.
Programming students might read the article and come away with the idea that a deterministic algorithm can generate random numbers.
This is like the sometimes-heard claim that "Our new error-detecting algorithm discovers whether a submitted program contains errors that might make it stop." Same problem -- wrong as written, but no qualifiers.
- tialaramex 1y agoBut this article converts another source of randomness, it doesn't care where the random bits came from. It will be just as happy with dice rolling, the PRNG from Commodore 64 BASIC, the built-in random numbers from a modern Intel CPU or "random" bits harvested by sampling keyboard input, it just takes the "random" bits and makes floating point values. So there's no reason to say it must be pseudo-random. In most practical applications it would be, but that doesn't seem essential AFAICT
- binarybitlist 1y agoSomething interesting happend, while reading... (Sry, none native English speaker, in German:) "Moment mal, sagst Du da gerade, mit aller höchster Wahrscheinlichkeit gibt es gar keine zufälligen Zahlen, die ein Computer generieren könne, da 'Bits' immer nur entweder den Wert '0' oder eben '1' annehmen können (und das auf inzwischen mehreren Layern, andere Geschichte...)?" let me try to explain it a little... 3 * (A = >0) adding... 1 * (A = 1) (..another topic yes...) ...but that what was wanted is "NONE" (of it...) or in other words 'maybe random'?? ?? :confused: