One part of the work of any researcher is to referee the work of other scientists in his field. This is a time-consuming and volunteer work, that is made with the idea of service for the community.
My main question: What are the standards of organization for the referee report of an article?
For instance, is a "general comment" section (mostly, explaining why the paper should be accepted or rejected) and a "specific comments" (in the form of a list of precise comments e.g. "Page $2$ line $4$: typo. It should read X instead of Y") enough? Or do people organize the review under specific sections like "Typos and writing", "Missing definitions", "Scope of the paper",etc...
I wish my report to be as useful as it could for the author, so I am inquiring about the best practices for referee reports.
PS: It may be a field-dependant question. I am mostly interested about standards in mathematics, but answers for other fields would be interesting too.