From Zotero to Mendeley

This is a New Year’s resolution that has actually come to pass. I have been a fan of Zotero for more than two years, a time in which Firefox was my browser of choice. By 6 months ago I had built a collection of about 800 references made so easy by the increasing capability of the Zotero Firefox extension. Zotero grew to about 2.5MB, a quarter the size of Firefox itself, as its feature list was extended. The constant updates began to be annoying especially as Firefox required restarting each time.

Like many others I used 2 or 3 different machines when building the Zotero collection, a problem since Zotero initially assumed a single, local database of references. Eventually I solved this problem by using Windows Live Mesh to synchronise the Zotero database and files between several machines (see previous post 1 and post 2).

I gorged heavily on Zotero’s excellent facility not only to store a web link reference but also to store the actual contents of a web page with all its images, styles and scripts intact. It was late in the piece that I discovered these files occupied about 700 MB on disk totalling over 70,000 mostly tiny files! No wonder Live Mesh struggled when my new netbook appeared – it took 12 hours or more to sync those files.

Zotero finally supported shared libraries and associated files but was limited to only 500 MB for free so I was stuck with Live Mesh. However another insidious development was taking place – I was becoming hooked on Google Chrome – and Zotero was only available on Firefox. In the end I turned to Mendeley which not only has shared libraries from the get go but also has desktop apps on all major platforms and a mobile presence on iOS.

Over the course of a couple of days I reduced my Zotero references down to well under 100. I discovered most references were captured web pages that:

  • had content likely to be out of date
  • could easily be found again by Google search
  • occupied large numbers of small files
  • only need their URL and title stored in a simple link database

It is clear the cost of storing captured web pages is only useful for transitory information that can later be recalled directly from local storage.

I therefore retained the bulk of the conventional references of journal papers, books, reports and similar for which Mendeley is primarily targeted. It was then very pleasing to see in Mendeley a direct import tool for Zotero as long as its database exists on the same machine on which the Mendeley desktop app is running. The only downside seemed to be that PDF files in Zotero references did not transfer automatically. Should you wish to continue using Zotero then this Mendeley tool will keep the two reference stores in sync.

In my case I bid Zotero a fond farewell even though multi-platform and smartphone support is promised.


About Michael Rees
Academic in IT interested in Web 2.0 and social media

2 Responses to From Zotero to Mendeley

  1. Avram Lyon says:

    While I wish Zotero didn’t make quite so many files in its web snapshots, I think you miss the point of the feature; you write that the snapshots “had content likely to be out of date” — this is actually the whole point of making the snapshots; they provide a copy of the page as it appeared when you consulted it. If the site content subsequently changes, you’ll need this snapshot. The point is to account for the transience of much web content.

  2. Michael Rees says:


    Yes, I take your point about web capture and indeed mentioned it’s usefulness for transitory information in my original post. When I reflected on my own Zotero entries, though, it was clear page capture was not useful.

    LIke you I agree the Zotero myriad files method for page capture is frustrating despite being an authentic copy. Other systems capture just the text and images or generate a PDF, something for Zotero to consider.

%d bloggers like this: