This question has been asked at least twice, but the first has an incorrect answer (as I point out in the comments) and the second only gives a "starter" instead of a solution.
I want to characterize the continuous functions $\mathbb R_\ell\to\mathbb R_\ell$, where $\mathbb R_\ell$ is the lower limit topology. Let's call these "lower limit continuous". They must of course also be continuous as functions $\mathbb R_\ell\to\mathbb R$, and these functions are known to be exactly the right-continuous functions. However, not all right-continuous functions are lower limit continuous. A counterexample is given by $f(x)=-x$, where the preimage of $[0,1)$ is the not open $(-1,0]$.
An initial conjecture I had is that lower limit continuous functions might be the "piecewise increasing" right-continuous functions. However, it turns out that right-continuous functions are not necessarily piecewise continuous, so the obvious definition of "piecewise increasing" doesn't work. A second guess was that these might be right-continuous functions where the domain can be partitioned into intervals $[a,b)$ where the function is increasing on each interval. This also turns out to be false, with a counterexample given by $$f(x)=\begin{cases}0,&x\le 0\\x-e^{\lfloor\ln x\rfloor},&x>0.\end{cases}$$ This prompts the main question and a bonus question:
- Is there any "nice" characterization of lower limit continuous functions?
- Are there any "nowhere increasing" lower limit continuous functions, as in functions that are not increasing in any open interval $(a,b)$?
Edit: take $q:\mathbb N\to\mathbb Q$ an arbitrary bijection and define $$g(x)=\sum_{k=0}^\infty 2^{-k}f(x+q(k)),$$ where $f$ is the previous counterexample. I think $g$ might be lower limit continuous and nowhere increasing, but I have no idea how to show this.