Crossref goodness of match knob

HLHJ

Apparently you can set a threshold for the Crossref metadata retrieval to reduce spurious results, and a good value is around 2.4. Could Zotero incorperate a control to set the threshold, with a 2.4 default?

aurimas

We only use CrossRef to retrieve metadata for DOIs, which are non-ambiguous. In the long run we may use CrossRef's search by author/title or the new metadata search engine to supplement existing metadata, in which case this may be a good suggestion, but that's not going to happen very soon.

aurimas

What problem are you experiencing with metadata retrieval, btw? I assume this is for retrieving metadata from PDFs, in which case it's probably because there is no detectable DOI and we end up querying Google scholar for metadata, which does produce occasional false positives. Are you experiencing this issue a lot?

HLHJ

Sorry for the totally ignorant suggestion, I didn't know Crossref wasn't used to get DOIs.

The most common metadata problem I get is searching by ISBN or something and getting a bunch of PMIDs, if I'm remembering correctly. It's a tad annoying because I have to delete multiple irrelevant items. My main problem with Google Scholar is incomplete metadata, like only having the first author, which I get with moderate frequency.

All in all, minor complaints. A togglable "one-at-a-time" setting on the ID lookup might sometimes be useful, but I can't think of any other easy fixes.

aurimas

I've seen this with ISBN import once, where the ISBN was missing a leading zero (hence invalid) and was separated by spaces instead of dashes or being a single number. Do you have an example? I think I agree with you though that importing without a confirmation is probably a bit annoying. We could add a dialog that displays some info about the import and then asks the user to confirm (it would generally be just pressing Enter). Dan, would you object?

HLHJ

I'm sorry, I don't have an example, but I'll try and remember to post the next one I get. It's only happened to me a few times.

I'm not sure I'd want a confirmation dialogue if there was only one thing being imported; lots of dialogues can also be a bit annoying, and I can delete one erroneous item easily enough.

But for lots of items, it would be great; I rarely look up list of IDs, and I've never done multiple lists in a row, so an importing-multiple-items confirmation dialogue would not be annoying. Of course, some people may use it differently. The dialogue could have a do-not-warn-me-again(for single/multiple imports) box.

Actually, though, an undo-import button might be more useful than a confirmation dialogue. I could also use it when I've just accidentally nabbed duplicate or inferior metadata. Is this much harder to implement?

aurimas

Undo import is just a delete, unless you're talking about multiple references, in which case this would involve a good amount of work.

Yeah, maybe just a dialog for multiple, since we do focus the newly imported item.

HLHJ

Plain delete for a single item and a confirmation dialogue for multiple items sounds great; thinking it over, multi-item imports sometimes take so long that an undo would not be prefereable to a confirmation box with a cancel button.

dstillman

Plain delete for a single item and a confirmation dialogue for multiple items sounds great

Yeah, that makes sense to me — and probably with checkboxes so that you can uncheck any that are wrong before continuing (since the majority might be correct).

HLHJ

Thank you to everyone who worked on this feature. It works well and saves a lot of time and frustration.