From time to time I receive papers to peer review that, in my view, fail to attain an acceptable level of writing and text formatting. Examples of usual issues are: common erratum, poor use of English, bad formatted equations, inconsistent use of punctuation marks, spacings, citations, references and so on. Bonus track is when the submission has not been processed in LaTeX and it contains poorly written equations (for the record my field is Mathematics). I do not consider myself an expert in any of these points, but I think everybody know how to reach a decent level, of course if you are willing to invest some time on it.
While in certain situations there are interesting scientific contributions to present, it is certainly hard to focus on them with so many styling distractions around. What is worse, sometimes they are hard to decipher because of an inconsistent use of notation or because a misspelling introduces a key change in the meaning of a statement. In my opinion, these are issues that could have been solved by the authors with a careful revision of the manuscript prior to submission.
For the sake of argument, let's assume a paper as described before, with a legitimate scientific contribution and let's keep apart from the question the scientific part of the report. When I face the review of such a paper, I often have the next dilemma: should I point the authors to all the writing/styling/formatting issues? Or should I briefly mention that there are serious non-scientific issues and encourage the authors to fix them? The two opposite arguments supporting each approach are:
On one hand, I feel that is my responsibility as a reviewer to ensure that the paper is going to have the best scientific/presentation quality possible after the reviewing process. And that includes helping as much as I can in both points.
On the other hand, if the authors should have taken care of all the non-scientific issues, I wonder why should I invest so much time pointing towards their improvement. In some sense, I feel they are being a little disrespectful with the reviewers and readers to submit a non-polished manuscript. In addition, maybe this kind of work could be done by the people in charge of the article proofs (assuming that the journal has them and that they do a good job...)
My current approach is to ask for a major revision and reply with a very detailed list of all the non-scientific issues I found, but I wonder whether I am overdoing my referee's duties. In addition, this turns to be also quite time-consuming.
EDIT: I am looking for an answer that involves a discussion on what are the reviewer's duties and limits regarding writing/styling. This is only partially addressed in this question.