It is unclear whether JOSS also checks whether the software is original, in the sense that it doesn’t re-invent a wheel. One of the goals of peer reviews of scientific papers is that the author is the first one to discover a result and puts a time stamp on it (see above). Does the journal allow any research software or does it help to avoid re-inventing the wheel? I miss this aspect about originality (see also section 4 on review criteria where it seems to be hidden in statement of need).
This section is a bit unclear on what is actually the target. On the one hand it is ‘ just’ advertising, on the other hand it mentions detailed code review.
If only high level functioning is reviewed, what is there actually to review in JOSS. How is quality and originality defined? So, how to determine impact and quality as peer review of scientific papers generally does? Does the code review have this role? Does this include coding standards, bench marking, checking for correctness of scientific result (in addition to code tests).