The Quantum Pontiff

Zotero 1.5 Beta and More

Zotero, a Firefox extension for managing research sources, has announced the release of Zotero 1.5 beta. I’ve heard good things from those who use Zotero. This major update adds some nifty synchronizing and automatic backup. The next step after this for Zotero, I believe, is adding sharing capabilities.

By the way does anyone know what happened to arXiv on you harddrive? It’s not been updated in over a year, which is a shame. Personally I find the arXiv’s lack of publicity accessible methods for obtaining the full text kind of a bummer. There is so much fun you could have given the ability to have the arXiv on your own local system. Sure there are bandwidth issues, but I’ve been hanging around long enough with computer scientists to know that there are serious good solutions to these problems as well.

Comments

  1. #1 Matt Leifer
    February 25, 2009

    I too have been looking at Zotero. It seems that there is a profusion of reference managers out there at the moment, and it’s difficult to decide which one to use. It is even harder to decide because import/export functionality is rather weak in some of these products, so you can easily end up locked in to one of them. Of the available choices, only Zotero and Mendeley are offering sharing/social facilities along with a comprehensive way of organizing a local repository. Unfortunately, I could not import an entire directory with lots of subdirectories of pdfs into Mendeley, so this renders it next to useless at the moment. I’m certainly not going to bother importing the files one-by-one. Also, Mendeley is modeled on Last.fm, which I absolutely hate, so probably is not the one for me. I also have the feeling that Zotero won’t let me import my pdfs easily. I can import the references easily via bibtex, but the file associations seem to be more difficult. Again, I’m not prepared to go through my files one-by-one to do this.

    At the moment, I am using the Mac OSX program Papers, which has a lot of great features, but is a bit bloated. Like many people, I was initially attracted to it due to the fact that it has a lot of eye candy and looks a lot like the sort of app that Apple would build. It models itself on iTunes, so perhaps the bloat not too surprising. My main complaints with this software are:

    - They have been promising an auto-completion of metadata feature that works like Gracenote in iTunes for years, but it is still not there yet.

    - It treats everything as if it were a “paper”, i.e. there are not separate content types for lecture notes, books, websites, preprints, etc. You can get around this by having a scheme to use the metadata fields for papers appropriately for different content types, but you have to remember what you did in order to make things consistent. Also, since everything is a “paper”, this makes import and export lousy. In particular, if you import a bibtex file and then export to bitex, you won’t get back the same thing that you started with.

    - You can’t associate more than one pdf with the same record. If I have an arXiv eprint and then later on I get a copy of the published paper then I don’t want to have two records. Instead, I want both files to be associated with the same record and I want there to be an eprint field where I can put the arXiv URL in addition to having the journal metadata. When I export to bibtex, I want all this data to be part of the same record. To be fair, I don’t think any of the reference managers have this feature yet, but it would be a deal-maker for me if someone implements it.

    I’m thinking of going back to Bibdesk, because its native format is bibtex, so I know it won’t screw up all of my old carefully hand-written bibtex files. Also, they have recently introduced the ability to associate a pdf file with a record, so it might not be too bad as a general purpose reference manager.

  2. #2 Matt Leifer
    February 25, 2009

    About arXiv on your hard drive. With the API, it should be fairly simple to write a script to download everything yourself. It would probably take a long time to run, since you’d have to ensure long enough gaps between API queries so that the arXiv doesn’t think you are a robot.

  3. #3 Dave Bacon
    February 25, 2009

    Yeah one could certainly do that (even using the OAI interface as well.) But what is really needed to make the whole thing useful is a distributed system so that the bandwidth requirements for others joining isn’t placed upon the arXiv site. One advantage of bittorrent is exactly this. However what one really wants is a system like bittorrent but with an ever updating file (I want my arXiv on my hard drive exactly in sync :) )

  4. #4 Matt Leifer
    February 25, 2009

    BTW, I just tried Zotero again and the import issues are just as I feared, i.e. it will import bibtex easily, but then you’ve got to link to the pdf files one at a time. No thank you. Looks like my library is stuck in Papers for the time being.

  5. #5 Heraclides
    February 26, 2009

    I tried Zotero some time ago and gave up on it, mainly for several reasons back then (from memory, excuse me if one or two of these are (now) wrong):

    1. No ability to start with a PDF, have Z. find the DOI in the PDF and work off that (BibDesk has apparently just added this).

    2. Linking to PDFs manually essentially ruined a large part of the idea for me.

    3. Not able to generate bibliographies for a manuscript.

    4. Unbelievably slow beyond a handful of references.

    I’ve moved to BibDesk since, which surprised me with how much it now can do. A few things still need a little polishing, e.g. the template editor is lovely idea, but needs to be a little clearer to first-time users as how to use it.

  6. #6 Victor
    February 26, 2009

    Hi Matt, hi Heraclides,

    based on what you said, maybe Mendeley would be interesting for you after all. Matt, even with your dislike for Last.fm it seems to have a couple of features you’re looking for.

    Mendeley’s desktop version does have metadata auto-complete, and it lets you attach multiple files to one record. You can also differentiate between document types.

    Heraclides, Mendeley starts out with the PDF – it lets you import folders and subdirectories (and then monitor these for new files continously), and finds and queries DOIs automatically. The next release (due next week) will do the same with arXiv IDs. If neither a DOI or an arXiv ID are found, it will try to “guess” the correct metadata based on the full-text.

    Besides, it also extracts the cited references from a document and indexes the full-text for quick searching. It then adds the social features, i.e. sharing selections of papers with colleagues, and it lets you sync your library across multiple computers.

    BibTeX Import/Export is supported as well, and we have a couple of neat features planned there, too. Essentially, whenever you update your research paper library, your BibTeX files will be updated automatically to match your library (but that won’t be in next week’s release yet).

    Best,
    Victor

    P.S. As an aside, our funding round was announced yesterday, so at least you know we’ll stick around for a while: http://www.techcrunch.com/2009/02/25/mendeley-snags-2-million-in-early-stage-funding-for-research-paper-management-tool/

  7. #7 Eric Lund
    February 26, 2009

    @Matt: At the moment, I am using the Mac OSX program Papers

    I tried that program a year or so ago, on the recommendation of somebody at ScienceBlogs, and very quickly abandoned it. I didn’t use it long enough to encounter the problems you cite, because I quickly ran into two problems I considered unforgivable:

    1. The program insists on calling its folder “Papers”, and its default location is in the Documents folder. The problem is that I (and presumably also many other scientists) already have a folder called “Papers” in my Documents folder, for the obvious reason that as a practicing scientist I occasionally write papers. There is no evidence that the authors of the program considered that possibility.

    2. When a paper has more than five authors or so (this is common in many areas of research including my own field), there is a risk that Papers would attribute the paper to the wrong first author. I observed this failure mode when searching for a particular author whose name was buried somewhere deep in the author list of the affected papers. This was the fault that caused me to abandon Papers–for fields like mine where papers are normally cited by author and year (e.g., Fulano et al., 2009), it is absolutely essential to get the first author’s name right every time.

    That was a pity, because my bibliographical database is the last thing I am using AppleWorks for (I couldn’t use Bento because compatibility issues prevented me from upgrading to Leopard, and FileMaker would be overkill in my case). I also, in parallel, keep a TeX file of papers I have actually cited, from which I cut and paste my references (but this technique requires hand reformatting if I am submitting to a journal other than the one for which it was designed).

    I’m a Safari user, so I’m not sure Zotero could work for me–and I have a low tolerance for commercial software that doesn’t work, so given what Heraclides reports I’m not eager to try.

  8. #8 Rick
    March 2, 2009

    BTW, I just tried Zotero again and the import issues are just as I feared, i.e. it will import bibtex easily, but then you’ve got to link to the pdf files one at a time.

    If your BibTeX has fields called ‘pdf’ that contain absolute paths to your PDFs, zotero will then import those PDF files & you will not have to individually link your files.

  9. #9 Rick
    March 2, 2009

    @Heraclides

    1. No ability to start with a PDF, have Z. find the DOI in the PDF and work off that (BibDesk has apparently just added this).

    Assuming the PDF has full text, Zotero can often locate the paper online & use online metadata.

    2. Linking to PDFs manually essentially ruined a large part of the idea for me.

    As above, this isn’t the case: you just need a specially-formatted BibTeX file (there is no single convention for how file links are used in BibTeX)

    3. Not able to generate bibliographies for a manuscript.

    Zotero can certainly do this.

  10. #10 Rick
    March 2, 2009

    @Eric Lund

    I’m a Safari user, so I’m not sure Zotero could work for me–and I have a low tolerance for commercial software that doesn’t work, so given what Heraclides reports I’m not eager to try.

    Zotero is only for firefox (and some firefox derivatives) & will not work with Safari. However, neither Firefox nor Zotero are commercial; they are both free/open source. Firefox+Zotero together have a smaller footprint than EndNote alone & so they may certainly still be worth a try even if Firefox isn’t your primary browser.

The site is currently under maintenance and will be back shortly. New comments have been disabled during this time, please check back soon.