Some years ago I came up with a weird idea which I think was of my own, although it is not sensible to make such assertions.
The idea was to define a function $f: \mathbb{N} \rightarrow \mathbb{N}$ which imitates the behaviour of the derivative of a product, i.e. (uv)' = u'v + uv'. Using the prime notation, I define the derivative of a natural number $n$ whose primary decomposition is $\; n = \prod_{i = 1}^{i = r} p_i^{\alpha_i}$ as $$n' := \sum _{j = 1}^{j = r} \alpha_j p_j^{\alpha_j - 1}\prod_{i \neq j}^{i = r} p_i^{\alpha_i} = n (\sum_{i=1}^{i = r} \frac{\alpha_i}{p_i}).$$
So I somehow applied the rule of the derivative of a multiple product, hence $(nm)' = n'm + mn'$ follows. Some particular cases are $1' = 0$ and $p' = 1$ if $p$ is prime.
Although I have never found useful this "natural derivative", I got interested on it. Here are some nice facts about the natural derivative:
- Solutions of $n' = n$ are $n = p ^p,$ where $p$ is prime. It has an air of the exponential $e^x$.
- The "average value" of $n'$ is $n \times \sum_{p \in \mathbb{P} }\frac{1}{p(p-1)}$
- One can reformulate the Goldbach conjecture as: All even numbers greater than $2$ are the image of a pseudoprime number (i.e. $pq$, where $p$ and $q$ are prime).
- It can be easily extended to $\mathbb{Q}$.
So now I have introduced to you this marvellous function, I would like to ask you:
Did I invented it or has it been used before?
In the second case: In which context does it appear, and how much is known about it?
Thanks a bunch for your answers.
