0

I was wondering why for example $\dfrac{1111_b \cdot 111_b}{11_b \cdot 1_b} = 11211_b$. Is there a good explanation for this and is there a name for products like this which have constant digits value independent of base? It also seems to hold with $\dfrac{11111_b \cdot 1111_b \cdot 111_b}{111_b \cdot 11_b \cdot 1_b}$ and other numbers as well.

Here is my general question:

Prove that $$R(m,k) = \frac{ 1^{(m)}_b \cdot 1^{(m-1)}_b\cdots 1^{(m-k+1)}_b}{1^{(1)}_b \cdot 1^{(2)}_b\cdots 1^{(k)}_b} = c_b$$ where $c$ is a constant and $b>2$.

Jyrki Lahtonen
  • 133,153
Puzzled417
  • 6,956

1 Answers1

1

This is because multiplication in a number system becomes a special type of convolution with carry. If the digits involved are small enough compared to the base used, then carry will never happen and we can calculate expressions like yours for example as a sequence of polynomial multiplications and divisions.

For example the first one: $1111_b$ can first be divided by $11_b$ to produce $101_b$, then the multiplication will be $111_b + 100_b\cdot 111_b = 11211_b$

mathreadler
  • 25,824
  • In fact, the generalization is in terms of repunits. Let $$R(m,k) = \frac{ 1^{(m)}_b \cdot 1^{(m-1)}_b\cdots 1^{(m-k+1)}_b}{1^{(1)}_b \cdot 1^{(2)}_b\cdots 1^{(k)}_b} = c_b$$ where $c$ is a constant. – Puzzled417 Jun 25 '16 at 13:14
  • Yes but it will work for arbitrary polynomials if we ensure the coefficients are not too large. Arbitrary polynomials are more general than products and quotients of repunits. – mathreadler Jun 25 '16 at 13:18
  • Can you prove it in the case of the repunits? – Puzzled417 Jun 25 '16 at 13:23
  • 1
    That was not your original question. – mathreadler Jun 25 '16 at 13:41