General > General Technical Chat
Rational, Irrational, and Integer. The correct definition.
<< < (11/12) > >>
Nominal Animal:

--- Quote from: CatalinaWOW on December 22, 2022, 01:03:57 am ---And writing an expression root(2) exactly and correctly defines one of these numbers.  But if you generated a histogram with this expression what happened in the machine was that an approximation was created and compared with the data.   Unless the data was all symbolically generated.
--- End quote ---
It was all symbolically generated, yes.  That's what "representation" as opposed to "approximation" means, at its core.


--- Quote from: CatalinaWOW on December 22, 2022, 01:03:57 am ---The fact remains that the only numbers that computers can deal with are the rationals
--- End quote ---
No, that's my point.  Any numerical object you can represent in arithmetic operations, you can use in a computer.

For example, consider the case of representing numbers using format \$S\cdot M^{X/F}\$, where \$S = \pm 1\$, \$M\$ and \$F\$ are unsigned integers, and \$X\$ is a signed integer.  You can represent many irrational numbers exactly with this form, although the algebraic operations you can do are a bit limited (only multiplication and division being trivial and exact).  An approximation is only needed when e.g. a human wants to see what the number is in decimal.

Almost all the computer hardware we use now does use integers, fixed-point integers (\$M/B\$, with \$M\$ a signed integer and \$B\$ an unsigned integer constant), and floating-point numbers (\$M\cdot B^X\$, with \$M\$ a signed integer, \$B\$ an unsigned integer constant most typically 2, and \$X\$ a signed integer).  These are all very efficient and effectivee representations for our typical needs –– that's why they're popular now, instead of the other possibilities! ––, but they are not the only ones, and it is a grave error to think of them as somehow a "requirement" or "definition".
TimFox:
A late reply to Nominal Animal's post above about my parable:

"(It is extra funny when one happens to know that you only need 206 bits to express the size of the observable universe in units of Planck length.  If the lines were half a meter apart, there would be fewer than 2³² atoms in the lattice between the marks, so you might be able to store the first four ASCII characters that way.  Apologies to TimFox for me trying to explain the joke..  ::))"

Hamlet:  "There are more things in heaven and earth, Horatio, / Than are dreamt of in your philosophy."

There is a whole lot more data extant than the extent of the universe.  There are unreliable estimates of the data content of the holdings (digitized or not) of the Library of Congress, but they range up to the petabytes, or roughly 253 bits.
SiliconWizard:
The Hamlet quote is great.
Nominal Animal:

--- Quote from: TimFox on December 22, 2022, 08:13:50 pm ---There is a whole lot more data extant than the extent of the universe.  There are unreliable estimates of the data content of the holdings (digitized or not) of the Library of Congress, but they range up to the petabytes, or roughly 253 bits.
--- End quote ---
Yup, I only meant to illustrate how little data one can encode in a single length.

If I try to consider how much information the universe contains, I get genuinely scared/overwhelmed.  :-[

For illustration, consider a cubic meter of iron, about 7874 kg, containing about 8.5×1028 ≃ 296 iron atoms.  Iron has four stable isotopes, so in theory we could encode two bits per atom just by selecting the isotope, so that cubic meter of iron would then encode 297 bits of information.

How much is that? Well, if you unraveled that cubic meter into a monomolecular thread, just one atom wide, and you could read it at light speed, your bandwidth would be about 262 bits per second, some 500 Libraries of Congress per second.  It would take about 1088 years to read the entire iron cube.

About 30% of Earth is iron, or about 269 cubic meters of it.  That's just on Earth, mind you.  Since iron is the end product of both fusion and fission chains, there is a lot of it in the universe.  And this is just a single information storage method with zero compression, just picking between four well-known stable iron isotopes.  There are much more dense methods; I just picked iron because it is utterly stable, easy to grasp, and Stephen Baxter wrote a story about it.
SiliconWizard:
I know I shouldn't add anything more here and it's not unlikely that the OP is just yanking our chain. With that said...

Universe is thought to be infinite, although we don't know that for sure, nor what this would really mean. But until we can prove it finite, it looks useless to me to try and estimate how many "elementary" particles it contains. If we have even an idea of the smallest ones, which we probably don't yet.

Secondly, to me the only real crucial point about "numbers" is whether, precisely, you admit infinity or not. (Which admittedly can itself be a philosophical matter since we have no means of grasping the reality of it.)

Once you admit infinity, pretty much any number can be derived from that. If you have integers, basic arithmetic operations and infinity, then you have rational numbers, irrational numbers (which can be expressed as infinite sums of rational numbers), complex numbers, which are just an extension of that all, and higher-dimension numbers, and so on. Just a thought. So do you admit infinity?
Navigation
Message Index
Next page
Previous page
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod