Both changes are likely to occur in the future. No matter the detection algorithm, there will be false positives that would need to be marked as non-duplicates. There is already a mechanism in the database structure that could be used for this (the "Related Items" feature could be expanded to support multiple types of relations), so adding non-duplicate marking is definitely feasible.
Great! That sounds good, it's clever if different relations can be set into "related items"-type of fields. I love Zotero, I hope to see these changes soonish!
Also see related topic here — I understand marking as non-duplicate is non-trivial but it's not clear it has advanced beyond the stage of being planned.
Same problem here. I switched to Zotero from Mendeley because of sync problems, but there was a function to mark records as non-duplicates. I have about 4000 references, many of them are from older sources and have no DOI. What can I do with this?
Anything new about whitelisting false positive duplicates? For me, it is really annoying and prevents me from de-duplicating the real duplicates because it takes so much time to take a look at all the false positives again and again in search for the real duplicates.
No news as far as I know. Here's a duplicate thread (hah) on this issue in which I periodically note how many false positives are littering my 'duplicate items' tab; 230 at last count.
Upgrade Storage
I love Zotero, I hope to see these changes soonish!