I'm reading a text which tries to prove that $$\lim_{x\to a} (f+g)=\lim_{x\to a}f+\lim_{x\to a}g$$
the first step goes like this "Let $\epsilon >0$, then because $\lim_{x\to a}f=k$ and $ \lim_{x\to a}g=L $ there is a $\delta_1>0$ and $\delta_2>0$, such that $$\vert f - K\vert<\frac{\epsilon}{2}$$
$$\vert g - L\vert<\frac{\epsilon}{2}$$
whenever $0<\vert x-a\vert \ <\delta_1$ and $0<\vert x-a\vert \ <\delta_2$"
Why is the distance to the limit not less than $\epsilon$? Why is it written as $\frac{\epsilon}{2}$, what is the rationale behind this?
This is the rationale.
We arbitrarily pick an $\epsilon > 0$.
We want to find a $\delta$ where where $|x-a|< \delta$ will mean that $|(f(x)+g(x)) -(K+L)| < \epsilon$. how will we do that?
Well, $|(f(x)+g(x)) -(K+L)|=|(f(x)-K)+(g(x)-L)| \le |f(x)-K| + |g(x)-L|$. SO if we can prove that
$|f(x)-K| + |g(x)-L| < \epsilon$ we will be done.
Now $f(x)\to K$ so for any $\overline \epsilon > 0$ we can find a $\delta_1$ so that $|x-a|< \delta_1$ implies $|f(x)-K| < \overline \epsilon$.
And $g(x)\to L$ so for any $\epsilon' > 0$ we can find a $\delta_2$ so that $|x-a| < \delta_2$ implies $|g(x)-L|< \epsilon'$.
So if $|x-a| < \min(\delta_1, \delta_2)$ the we would have
$|f(x)-K| + |g(x)-L| < \overline \epsilon + \epsilon'$.
But we WANT $|f(x)-K| + |g(x)-L| < \epsilon$.
So if we can have $\overline \epsilon + \epsilon' = \epsilon$ we will be done.
... that is our proof would be.....
Now the only thing left is to figure out for any arbitrary positive $\epsilon$ how do we find positive $\overline \epsilon, \epsilon'$ so that $\overline\epsilon + \epsilon' = \epsilon$?
And... well,.... if $\overline \epsilon = \frac {\epsilon}2$ and $\epsilon' = \frac {\epsilon}2$ then $\overline \epsilon + \epsilon' =\epsilon$ is a good choice. We could have chosen any other pair though.