I’ve really, really struggled with the choice of citation managers throughout my PhD. I’ve ended up going with Zotero, because I know my data isn’t locked in (FOSS, works offline without a login), and because it supports the particularly weird citation management in my field via .cls files.
However, the UX in Zotero is far from ideal and I find it really clunky to manage different collections and quickly find relevant documents. I previously used Papers 2, which had an excellent UX (although was limited in other ways). To take one example of this: when you assigned a colour to an item in papers 2, the background of the whole row changed colour. It made visually identifying particular items very easy - I used to have book reviews purple, for example. In Zotero, this is done through tags, where you can then assign colours to particular tags - but you just get a small coloured square at one end of the item in the list, which is almost invisible. Similarly, editing tags requires switching a pane and is very clunky.
I feel there is a real lack of a good software in this space, especially for Macs - Bookends seems to be the closest. I have dabbled with this, which I think is closest to the “old” Papers in terms of a very committed developer and passionate community, but I’m not a huge fan of the interface and when I tried it, found certain things somewhat counter-intuitive. I haven’t yet tried really committing to it for a project and perhaps it would grow on me with time.
Too many of these kinds of reference managers have problems, in my opinion:
- Some try to monetise via ongoing subscriptions - when you are building a life-time collection of data and research materials, this not good at all (and often import/export is limited or incomplete), as you are then tied into paying a subscription indefinitely.
- Some rely on online components, such as for logging in - some friends discovered this to their cost when Mendeley went down a few hours before a major conference deadline and they couldn’t access their references/update their papers.
- Some, especially those owned by large publishing companies, are definitely profiling you and your data.
- Some are very expensive (EndNote).
- UX is often poor, at best, though this varies a lot.
- Many are lacking features which some or many users would need, or have other limitations.
In the latter two items I would include, for example:
- BibTeX export
- a good companion iPad app (or easy standards-based way of accessing PDFs from tablets - some apps store these in a companion database that require workarounds to ‘export’ and re ‘import’ them)
- accurate ‘fire and forget’ downloading of metadata when you drag PDFs to the app
- incomplete metadata fields (very important in some fields - I had to give up on Papers 2 when I published some computing conference papers, because not all required citation fields were shown in the app even though they still existed!)
- ease of editing metadata. Papers 2 was excellent about this, Papers 3 required switching panes completely and was much more clunky. Zotero is middling in this regard - it doesn’t obey standard keyboard shortcuts, and there are a lot of ‘nice to have’ UX features that were in Papers 2 eight years ago which it doesn’t have.
And much more… I did once write a list of features an ideal app would have, once I finish the PhD I’ll dig it out.