I’ve been a big fan of bibliography software and I’ve written before about how much I like BibDesk for this role. That said, my workflow just wasn’t complete until I gave Zotero another try recently. The software has come a long way since I first looked at it, and now I am happy to say that it is my primary tool for tracking resources while I’m searching through articles. Two features finally sold me on zotero: the meta-data is scraped correctly (identified and stored) from the web resource in all of the journals I have tried, and I can easily sync my library and my stored PDF files between multiple computers. File sync is free through zotero.org (up to 100 MB) but I have a WebDAV server at work so I use that for unlimited storage. As a major bonus, I also like the feature that will automatically download and link the full-text PDF file (if you have access to it). These three things combine to make a great tool.
The one thing I discovered that may help you find success is that zotero creates a button in the firefox address bar whenever it can save the article metadata. This button seems to do the right thing when you want to save a journal article with all the metadata and a link to the original abstract page. I had different (and sometimes wrong) results when I tried to “Create New Item from Current Page” using the button on Zotero’s toolbar. This may have been addressed in the latest version, but I just use the addressbar button (since you can click that without even opening the zotero application pane).
I haven’t had to write a paper yet after collecting these citations, but I will continue to use BibDesk for managing bibtex files since I write in LaTeX. Fortunately the intergration between zotero and bibdesk is fairly straightforward, especially with George MacKerron’s Zot2Bib.