Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Are you sure? Isn't the average calculated by dividing by n, an integer?

Edit: I mean, you're right that nearly every number is irrational. But I think averages are going to be some of the tiny fraction of numbers that aren't.



> Edit: I mean, you're right that nearly every number is irrational.

The set of rational number is a “null set”: https://en.wikipedia.org/wiki/Null_set


Only if you take an average of rational numbers. Like e/2 is still irrational.


While in theory you could measure sleeping time in fractions of 2 pi, I'd guess that you're using rational numbers in the actual calculation anyways.


Why would sleeping time be a fraction of 2pi?

> I'd guess that you're using rational numbers in the actual calculation anyways.

Well, actually one uses floating point arithmetic, which isn't rational numbers either (as shown by the classic example a = 1/3, b = 3*a, then b != a)


Base 12 being 3-smooth, any number with 2 or 3 as a factor has a reciprocal with a terminating expansion.

base 60 is 5-smooth, so any number with 2,3,5 as a factor has a terminating expansion.

Just as SI has to add in deg min sec for fields like astronomy.

As the practical numbers are quite dense up to 60, they could divide by multiplication of the reciprocal for many more numbers.

Floating point is more complicated than just the base as the radix also matters. C(++) finally got decimal radix support this year in the standards and IBM has had decimal floats for a long time.

FFT and encryption often use mixed radix despite being a binary base too.

It is a far more complicated subject than it appears on the surface.

But there are problems that aren't easily solvable in base 10. The degrees of a circle are an example and why navigation uses the nautical mile, where 1 nautical mile= 1 minute of latitude is an example.

12,60,360 are Superior highly composite numbers and 12 is the smallest 3-smooth and 60 is the smallest 5-smooth.

This also means that with using 360 degrees one can divide a circle or semicircle in 12 sections with just a square, 345 triangle and equalatrral triangle. Where decimal or even radians requires the square root of 2, pi, etc...

I am a fan of universal units of measurement, but had they been base 12 it would have been better IMHO. SI could be more broadly adopted if it has been base 12.


Base 6 is better than base 12.

https://www.seximal.net/


Pretty light on details, but if it works for you that is great.

Quartiles and hand counting would be my argument for 12 but am probably biased based on familiarity.


Seximal is actually great for hand counting. You can use fingers the classical way for one digit per hand or base 6² compression for two digits per hand, allowing you to count up to 1296 with two hands.


I mean, we could be measuring time in radians. We already subdivide degrees into minutes and seconds, might as well call the day 2 pi radians.

We sleep for 2/3 pi radians of the day in that case.


The whole argument started when somebody above claimed that we probably do not sleep exactly 8 hours, but some weird irrational number. That problem persists even if you measure in radians...


Do the physicist thing and just define all your constants to be one.

"How long is a day? 1 day."

"I slept for .37 days."

This nicely unifies things with the tao manifesto[1] people since 1 day = 1 turn = 1 tau = 2 * pi radians and reinforces the periodic nature of these things.

[1]: https://tauday.com/tau-manifesto


Yes, but in reality, you slept .37827387654329617345987613475130465019873458612386459012374587203475. . . days


Doesn't that depend on whether time is quantized?


Yes, but with no viable theories of a minimal unit, I'll assume that no such thing exists until given proof otherwise.



> Why would sleeping time be a fraction of 2pi?

The rotation of a hand of the clock, expressed as the arc length of the curve of its endpoint :)

A sleeping time of 0 is always a problem, though.


But why would the second factor be a rational?


Actually averages are even more likely to be irrational than ordinary numbers.

It only takes a single irrational input to make the average irrational.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: