Hi All! I hope you don't mind me barging in on this thread. I'm the academic community liaison for Mendeley, so I can answer some of the questions people are asking.
The idea behind Mendeley is to get you up and running as soon as possible with a minimum of data entry, so you can just point it at the directory that has your PDFs and it will extract the citation information for you, optionally renaming and organizing the files, too. As timuralp noted, getting the metadata right is challenging, but we have several different ways of getting the right info - Extraction from the XMP metadata, Google Scholar search on the title, or lookup of the DOI, PMID, or ArxivID from Crossref, Pubmed, or Arxiv.org. If all of these fail, you can manually enter the missing info. Like Branwyn says, it generally works pretty well. You can also fetch information directly from Zotero or from exported Endnote files.
In terms of the comparison to Zotero, Papers, and Endnote (of course I'm biased) but the major differences are:
Mendeley vs. Zotero - Mendeley is a standalone program, so you don't have to use Firefox for it to work and your data isn't subject to browser crashes.Mendeley vs. Endnote - Mendeley is free and designed using modern design principles. Little manual data entry is required and you can search the full text across all your papers. Mendeley also has annotation, sharing & discovery tools lacking in Endnote. [*]Mendeley vs. Papers - Mendeley is free and cross-platform, so you can collaborate with anyone, regardless of OS. Also, while Papers organizes files very nicely, Papers doesn't have an add in for Word and OpenOffice.
Here's a less biased version comparing all the major reference managers: http://blogs.nature....ence-management
Hope this was helpful! Please don't hesitate to get in touch (william.gunn@mendeley.com) if you have any followup questions.