To answer your last question, an errata is always (almost always?) written by the authors of the article, whereas a "Comments on..." article can be written by anyone. However, to write a "Comments on..." article, which in my experience are rare, you must have a substantial new idea. It's not enough to say "I found typos in the original paper".
If the original paper contains a proof that is incorrect, and you have a proof that is correct, that could be enough for a paper. In this case you probably also need to provide counterexamples to the original proof. Similarly, if you can explain why the original data analysis was seriously flawed, and you have correct data analysis, that might be a paper.