There are several questions on this site about errors in papers and even the errors-erratum tag. I'm specifically thinking of mathematical errors, where there is an unambiguous mistake, like a wrong sign or wrong physical constant. I have come across ~5 of these in the past month.
Is contacting the publisher really the best way to correct these? Sometimes papers are old and it seems unlikely that a new PDF will be generated. Would the academic research community find it useful to have one canonical site that made it very easy to submit errors and corrections. Then, when reading a paper, one could check the site with the DOI or title and see if that paper has an entry. I am considering making such a site, but wanted to check with other academics if it seems like a worthy endeavor, if a site like this already exists, or if there is a reason it does not exist.
I realize this is an opinion question. I know these types of questions get closed, but this is still the best site I can think of to ask the question, so hopefully I get some good replies before this gets downvoted.