COinS adds non-canonical url

kareliot

Whe using the COinS translator to add books to zotero from firefox (for example from this library catalogue [opac] http://www.gateway-bayern.de/) zotero (or the COinS translator) adds a URL from the database (something like https://opacplus.bsb-muenchen.de/search?isbn=0195644573&db=100). This should not happen, as the url does not point to the cannonical source.

Instead the translator should only add the snapshot to the url (which is added correctly)

dstillman

If you're referring to the URL field, I'm not seeing that. Can you provide exact steps to reproduce?

Oh, never mind. Reproduced with the second link you gave. We'll look into it.

dstillman

So, basically, that second page sets https://opacplus.bsb-muenchen.de/search?isbn=0195644573&db=100 as the rft_id, and if there's an HTTP(S) value for that field then Zotero assigns it to the URL field.

I'll let others who are more familiar with COinS comment on the merits of that, but it seems like correct behavior to me.

noksagt

Agreed. Zotero is doing the correct thing.

OPACplus has mis-encoded themselves as the referent rather than the referrer (rfr_id). They should fix things on their end.

I can't think of a generic way for Zotero to "know" that a site has bad data. It could work around bad data using site-specific translators, but I'm not sure if this data source is popular enough or if the defect is large enough to justify that.

kareliot

Thank you for looking into the matter and for your comments. I will try to contact the persons responsible for the implementation and inform them about the misbehaviour of OPACplus.

adamsmith

(@zuphilip and I have been talking about proper Zotero support for OPACplus, I think, but somehow never got around to it. Obviously that would be the proper solution for this)

zuphilip

Yes, this is still somewhere in my TODOs. The second page also provides a link to the MARC data *[@id=#marcLink] which maybe could be provided in a way that unapi translator recognizes it...