Wednesday, July 23, 2008

Zotero - bibliography management in firefox

For a while now I have been saving the academic papers I read on line on my laptop in folders by journal wih names like author_title.pdf. I have a little bit of automation: a bash script converts pdf to text, then spits the first few lines of the file, typically containing the title and authors, in to an html file with a link to the local copy of the file. The advantage is that it is minimal trouble. I look at a lot of papers on line and my memory is not so good, this is quite fast and does the job. Of course what I really want is this automated so with one click in my browser the pdf is saved and an entry saved in a database which can export to BibTeX for use in papers I write.


Well fortunately someone has done all of that. Zotero is a Firefox extension that scrapes bibliographic information from journal web pages, Google Scholar, ISI Web of Science, MathSci, etc

If the full text is available, and "automatically attach pdfs" setting selected in preferences it saves a copy of the paper in a directory that can be accessed from the database (or otherwise). references can be saved in a number of formats including of course BibTeX, but also RIS which is popular eg with Endnote.

My own university library, the John Ryland University Library at the University of Manchester has an OpenURL server called "FindItatJRUL", this can be linked to the Zotero database. To do this go to "preferences" OpenURL, custom and paste http://openurl.man.ac.uk/sfxlcl3 in the box. Goodness knows how this works. I just made an informed guess. But it does. My work for your library too.

I did try EndNote a while ago and dint find it very useful. In any case I do not use Windows or Macs, only Linux (Suse 10.3), and it did nor seem worth to trouble to run it in Wine or CrossoverOffice. Strangely though in Web of Knowledge clicking the "Export to EndNote RefMan or ProCite" icon, then ignoring the option to save a file, but then clicking on the little zotero icon in url box of Firefox does the trick. In Google Scholar clicking on "Import into Endnote" does the job with no fiddling.

It remains to be seen how well this works out. For example how easy it is to restore the database after say reinstalling Firefox and Zotero or other possible nasties, or if any other annoyances crop up. It does seem to be under active development though, so I expect bugs will be ironed out rather than get worse, and integration with popular databases will probably get better.

For MathsSciNet I found I had to use the main www.ams.org/mathscinet/ site rather than the german mirror I normally use. Also it fails for older MathRev records starting with zero until that is fixed. Of course off campus, instead of using the crappy University (ie cisco) VPN, I use "ssh -ND some-port me@somecomputeroncampus" (see also a better way to proxy" and then the FoxyProxy Firefox extension to route requests to places like MathSciNet that are not authenticated by "athens" or "shibboleth".