Until otherwise $e=\frac{\text{Domestic currency}}{\text{Foreign currency}}$.
The UIP condition can be written in the following form: Investing now in a domestic bond, must give the same return as investing in a foreign bond, which means, $1+i_t=e^{-1}_t\left(1+i^*_t\right)e^E_{t+1}$, implying
$$\frac{1+i_t}{1+i^*_t}=\frac{e^E_{t+1}}{e_{t}}=\frac{1}{1+\frac{e_{t}-e^E_{t+1}}{e^E_{t+1}}}$$ where $i_t$ is the domestic nominal interest rate at time t, $i^*_t$ is the foreign nominal interest rate at time t, $e^E_{t+1}$ is the expected exchange rate at time $t+1$.
Using the derivations presented here I get: $$i_t-i^*_t=-\frac{e_{t}-e^E_{t+1}}{e^E_{t+1}}$$
However, if we define the $e=\frac{\text{Foreign currency}}{\text{Domestic currency}}$, we get $1+i_t=e_t\left(1+i^*_t\right)\left(e^E_{t+1}\right)^{-1}$, and using an analogous reasoning to the above: $$i_t-i^*_t=-\frac{e^E_{t+1}-e_{t}}{e_{t}}$$
However, this wiki link gives(notice the missing minus sign):
$$i_t-i^*_t=\frac{e^E_{t+1}-e_{t}}{e_{t}}$$
It's not the first time I see this last(wiki) approximation being used.
What's the reasoning for this last approximation? are my approximations wrong? I would really like to know the reason for the minus signs to be missing.
Any help would be appreciated.