Are there plans/step plan for bootstrapping a non-test version of duniter?

Hello Arcurus, everything is going fine so far, @cgeek is finishing a really nice version which will clean and simplify a lot the core code :slight_smile:

Here are the explanations (in french) : Formule alternative pour le DU - #61 by Galuel, which is translated nicely by google :

To understand simply the proof of DUĞ = DU (t + 1) = DU (t) + c² (M / N) (t)

By definition, DU is the variation of M / N with: (M / N) (t + 1) = (M / N)

It is enough to multiply by c the terms of this equality, we obtain: (E): c * (M / N) (t + 1) = c * (M / N) )

The equality is also evident:

C (M / N) (t + 1) = DU (t + 1) (1)
C (M / N) (t) = DU (t) (2)
DU (t) = c * (M / N) (t) (3)

(E): DU (t + 1) = DU (t) + c² (M / N) (t) is replaced by (1), (2) and (3)

The advantage of this formulation is that it takes into account the variations of “N” while ensuring a high stability to the evolution of the DU (t), even in the case of a great shake, since if N decreases by a factor 1 / c (Factor 10 for c = 10%), the DU (t + 1) will only vary by a factor of 2, and if N increases by a factor 1 / c the DU (t + 1) will always remain slightly above DU (t).

This formulation is called “second order approximation”, whereas DUA, DUB and DUC are “first order approximations”.

The higher the order of an approximation, the more complicated the formulation, but the more correct it will be. However, it is also shown that if the calculation step is very fine, (the unit of time is “small”, or the “c” of calculation is “small”) plus a given approximation will be just as well.

We note that the graph of DUĞ smooths the variations with flexibility, whereas the approximations of the first order either do not take into account the variations of “N” or show “acute angles” when returning to normal.

Basically, it’s the same equation but at an higher order. It’s more precise and support better the evolution of members count in the community.

10% ! :slight_smile:
(1+10%)^(1/365) = 1.00026 → 0.026% a day

At the moment this is not something we imagine. We will try to have fixed parameters at first, but if it causes problems to the WoT, we will have to find a way to make parameters evolve. We will think about it if necessary when it happens I guess :slight_smile:

1 Like