Thanks for the exemplary bug report. One correction: This only happens "When some works have the same title and authors" and the same year and then searches for it. Therefore this really doesn't affect a lot of people. Personally I think it is bad style to publish two different papers using the exact same title in the same year. The above conditions are not the only place where confusion may arise. I think adding a suffix, like "extended version" or "with proofs" or similar is the better practice.

Yes, probably. The default grouping/sort order is by year and any entry that doesn't have a year field will be grouped under undefined.

By the way, 6000 entries is a *lot*! Do you really want your users to look through all that? You might be interested in using filters to provide different views.

I can delete posts, and sometimes have to when spammers show up. So if you ever need one deleted, just let me know.

Re. your question: they show up as different entries for me:

It was a caching issue. I loaded the URL once with &nocache=1 to force-update the cache, and not it's showing. How recently did you add these additional publications to your Zotero?

By the way, I recommend against using iframes for embedding.

Could you send me the link to your page?

Probably just a caching issue. Can you send me a link to your page?

Can you send me the link to your page? You probably just have some sort of library conflict.

The counting doesn't distinguish between users. It just counts clicks per publication, where publications are identified using their bibbase-id constructed from author names, title, and year (the last part of the url when you click the "link" link). So no, it doesn't reset ever.

We just count the number of times users have clicked one of the links listed in the publication's url fields, which are rendered as little icons in the default style, e.g., for PDF.

Thanks for reporting this. It has been fixed now.