decide on url_o introduction

[LukasWallrich]

We still have some original studies with non-DOI identifiers - need to decide what to do with them. Even if rarely used (so far), I would suggest to introduce a url_o field analogous to url_r?

Invalid DOI format (4 issues):
DOI does not match the expected pattern 10.NNNN/....

art00005 | 1 | 10.1007/s13164-018-0400-9: doi_o=art00005; doi_r=10.1007/s13164-018-0400-9
djaerv_2018 | 2 | 10.1098/rsos.231240: doi_o=djaerv_2018; doi_r=10.1098/rsos.231240
escholarship.orgucitem39d347bk | NA | 10.1098/rsos.231240: doi_o=escholarship.orgucitem39d347bk; doi_r=10.1098/rsos.231240
thekeep.eiu.edu/theses/4783 | 1 | https://thekeep.eiu.edu/theses/5025: doi_o=thekeep.eiu.edu/theses/4783; doi_r=NA

[LukasWallrich]

From earlier issue:

Do we need a non-DOI identifier? Maybe introduce url_o and then identify both _o and _r that do not have DOI by their URL? Or have an identifier column that can also be something else that is stable? At least having other values in a doi field is bad practice and makes validation harder ...