Over the last four years, I did what a normal person would call "impossible."
I would call it something else:
"the only way not to use my brain every second of my life."
I built modules out of mana.
Essentially—nodes that perform operations on two numbers.
Yes, you could call it a calculator.
I have an addition module.
Subtraction.
Multiplication.
Division.
All the basic operations.
Addition and subtraction work across a very wide range—not because I suddenly became a god, but because those two collapse into adders.
If I have a chain of N bits, I can add numbers up to 2^N.
In practice, my current limit is around a thousand bits.
A thousand adders.
Yes.
It sounds dramatic.
And I wouldn't believe myself either, if not for one "but":
mana grows with age.
control grows with practice.
Result: the range is enormous, and the whole thing computes in a couple seconds.
But of course, this world doesn't give gifts for free.
The main problem isn't what I _can_ do.
It's what it _costs_.
If I build a thousand-bit module, I won't have mana left for anything else.
So I made the most reasonable decision engineers usually make when resources are limited:
reuse.
I implemented multiplication and division "longhand"—as repeating steps that reuse addition/subtraction, shifts, and checks.
It's slower than "an ideal hardware circuit."
But cheaper.
And my world loves being cheap.
---
Now, the thing you, as an author, seem especially interested in.
Fractions.
And numbers like:
121.23 * 0.92
Yes.
At Level Thirty, the system suddenly does _this_—and smiles right into my face.
At Level Twenty-Nine, the problems were stuff like 12.2 * 8.3.
And at Level Thirty—121.23 * 0.92.
The jump is… impressive.
At first I thought it was a bug.
Then I remembered where I live.
There are no bugs here.
There are only "surprises."
---
Decimal fractions are their own special kind of pain.
Because my calculator thinks in binary.
And the world thinks in base ten.
And if you want to work with fractions properly, you have two paths:
1. store the fraction as numerator/denominator;
2. use floating point.
I chose the second.
Because the system feeds me decimal fractions, not nice clean 1/2.
So I took the idea behind IEEE 754.
Not the entire standard, just the core concept:
a number is stored as:
- sign (S)
- exponent (E)
- mantissa (M)
Meaning:
(-1)^S * 1.M * 2^(E)
And I built my own version.
ManaFloat.
I can expand its precision if I have enough mana.
I can shrink it if I want speed.
It's not "perfect."
But it works.
And that's what I need.
Most importantly: to work with fractions, I have to be able to convert them into binary.
And here the algorithm is simple, but infuriating.
Take the fractional part (for example, 0.92).
Multiply by 2.
Look at the integer part—that's the next bit after the binary point.
Keep the new fractional part and repeat.
Step by step.
0.92 * 2 = 1.84 → bit 1, remainder 0.84
0.84 * 2 = 1.68 → bit 1, remainder 0.68
0.68 * 2 = 1.36 → bit 1, remainder 0.36
…
And so on.
Until you get the precision you need.
And yes—sometimes a decimal fraction becomes an infinite repeating fraction in binary.
Which means precision is always limited.
So I added rounding and extra bits (guard/round/sticky style), so the errors stay small and predictable.
I don't like the word "error."
But I live in a world where error is the only way to make computation real.
My modules currently hold for about an hour before they start to spread and I have to rebuild them.
Sounds bad.
But compared to what I had before—this is massive progress.
An hour is already a "tool."
Not a "toy that breaks the moment you blink."
And now—here's the main point.
I'm turning twelve.
I hit Level Thirty a long time ago.
And for the first time in a long while, I feel like today something is going to change.
And honestly…
I really want to know what happens next.
---
Author's note: I've written about 40k words so far for this book; 9.2k of that is Chapters 1–9. That's obviously not War and Peace. I do have some writing experience, but not a ton, so please don't judge too harshly.
Right now I don't have a server, a Discord, or really any "place" where this can be discussed. And I genuinely want to understand: does it make sense to create something like that now—or is it better to wait until the book is at least a little popular?
I'm a programmer by profession. I _can_ explain the technical side in extreme detail. But at some point I started wondering: who is this actually interesting to? To care, you need a lot of context. I could easily write another ten chapters just to properly cover division and multiplication—for integers and for floats… but who's going to read that?
If someone is truly interested in how it all works, they can just google it, or even take a couple college lectures. It's not insanely difficult—just takes a lot of hours to break down.
So: I'd really love to hear your opinion—how do you want the story to develop next? I'll try to read all the comments… assuming there aren't too many :)