Mendeley: The future to managing science papers?

by Eli on March 22, 2010

This is another guest post by Sarah Kendrew (blog, Twitter) a postdoc in infrared astronomy at Leiden Observatory working on instrumentation for the European Extremely Large Telescope and the James Webb Space Telescope. She’s also contributed an article about Evernote.

Mendeley LogoStaying on top of the literature, even in a narrow field, is one of the biggest challenges we face in research today. Do you have an ever-growing pile of astro-ph papers on your desk you’ve meaning to read? Yeah, we all have that. In recent years a number of software packages and web applications have come on the market to help researchers organise their literature: Papers, Reference Manager, Jabref, and Zotero. Past AstroBetter posts have introduced Papers and discussed Papers vs. BibDesk. A recent addition that’s been getting good press lately is London-based Mendeley. Here I’d like to share my experience and thoughts about Mendeley as an astronomy tool.

Mendeley Screen ShotAt its simplest, Mendeley is a very handy piece of software to help organise and keep track of the papers you have on your hard drive. Via a desktop application you can create a literature library of papers which can sort, create collections, tag and make the papers searchable. The programme parses pdf files and automatically creates a record for each paper, straight from the pdf. This works really well for recent papers, but for those that are older than 5 years or so, you may need to edit some fields manually. On the upside, the programme is good at sensing it’s done a bad thing and will prompt you to verify the info when in doubt. One missing feature is the ability to parse postscript files, these have to be imported and edited manually too.

Mendeley can actually re-organise your mess of papers into an ordered directory structure, sorted according to whatever criterion you want – author, date, subject and so forth. Say, if you store your papers in C:/papers, you can tell Mendeley to “watch” that directory. Whenever you save a paper to that directory from e.g. ADS, Mendeley automatically pulls it into your library database. So you don’t have to manually create a record for each paper you save to your HD.

When I’m writing a paper, where having my literature in order is key, Mendeley really becomes most useful. There’s nothing worse than spending an hour trying to find a paper on my hard drive, only then to have forgotten what I wanted it for in the first place. I write my papers in LaTeX, as most of us do, and use BibTeX for easy referencing. In one of Mendeley’s latest upgrades, automatic BibTeX file creation for each individual collection was added. So my collection of papers on, say, star formation now has a BibTeX file connected to it that gets updated every time I add a paper to it. It’s fantastic! For those using MS Word (a necessary evil for us instrumentalists), there’s a nifty add-in for referencing and creating bibliographies.

We have only seen the tip of Mendeley’s iceberg and what it is all about. Mendeley tags itself as “Last.Fm for research”, a social networking site that gathers info on what scientists are reading and saying about literature. The uptake in the astronomy community isn’t really significant enough for the reading statistics to be meaningful yet, but the online presence has many advantages. Users’ libraries are connected to an online profile, and via the website they can connect to colleagues and friends, and discover what others are reading. This doesn’t mean that your contacts see your entire library – but you can share publications with them, and if you add your own publications to your profile they will be visible to your contacts. My whole library is synched to the website, including annotations – so even when I’m away from my own computer I can look up references. I can also add things to my library via the web, using a Mendeley bookmarklet or directly from Google Scholar, NASA ADS, or even astro-ph. That’s what those icons are for in the right sidebar – who knew? So I don’t need to have a desktop application running all the time to add papers to my library – next time I open it, it’ll get all the new stuff I added from my web profile.

One of my favourite features is the ability to create shared and public collections. Shared collections are for collaborations, everyone involved can contribute papers and annotate. I’ve actually not used shared collections yet myself but imagine them to be very useful if everyone in a group is willing. Public collections are, as the name suggests, entirely public. I recently created a public collection called Gravity, to accompany a blog post concerning a paper on a novel theory of gravity; in effect I created a reading list to illustrate what I was talking about, that I can add notes and comments to. Public collections have an RSS feed, so anyone can actually subscribe to a collection and stay up to date with what gets added. They’re also embeddable, so I can add them to a webpage. For those of us who teach, this could be a really awesome way of connecting with students, providing them with supplementary material in a very low-effort way.

I haven’t sampled all of Mendeley’s features but for me it’s proven to be a great tool to help me stay organised. I used to stress about sorting out my hard drive to organise my papers, which were nicely ordered during my PhD but are now a total mess. Now I let this piece of software do it all for me. The one thing it won’t do is send the information from the paper into your brain. There’s no app for that, only grad students/postdocs (delete as appropriate), so hang on to those for now.

Sarah Kendrew blogs about astronomy science and politics at SarahAskew (RSS, Twitter). She is also a postdoc in infrared astronomy at Leiden Observatory working on instrumentation for the European Extremely Large Telescope and the James Webb Space Telescope.

{ 15 comments… read them below or add one }

1 Sarah March 22, 2010 at 7:11 am

By the way, I’ve put more screenshots in my public Astrobetter notebook on Evernote, here. If you want to see what a public collection looks like on Mendeley, see it here.

Reply

2 Kelle March 22, 2010 at 8:13 am

I’m really excited about using shared collections for our Reading Group!

Reply

3 Sarah March 22, 2010 at 8:27 am

Oops link broken in previous comment. Public collection URL is:
http://www.mendeley.com/research-papers/collections/1078451/Gravity/

Reply

4 Tanya March 22, 2010 at 1:08 pm

This looks very promising. Thanks for sharing!

Reply

5 Jonathan Foster March 23, 2010 at 3:35 pm

After playing with the program for a little while I’ve found one thing that seems to break my mental conception of how this program should work.

I browse references on ADS of course. When I find the abstract I’m interested in, I can use the browser script to import that reference into Mendeley. This works fine, and produces a reference to the paper with a link to the ADS-abstract page, which is the most useful link to have. However, for getting the PDF, I have to link it by-hand in Mendeley. If I just import the same article’s PDF (downloaded from ApJ, say), then Mendeley will have two references to the paper, one associated with the PDF, and one associated with the ADS entry. I can find no way to tell the program to “merge” these two entries, despite having the same bibliographic information. The PDF-parsed entry is inferior to the ADS one, since it only links to the journal (IOP, in this case) homepage for the article.

Have others encountered this problem and/or found a solution that works for them?

Reply

6 Mr. Gunn March 24, 2010 at 8:21 pm

Johnathan, that’s not how it’s supposed it work. The PDF entry is usually associated with the citation, but in the cases where they aren’t you should be able to delete the PDF only entry and then drag and drop the PDF itself onto the citation entry. Another way to do it is to do a title lookup on the PDF entry, but really, Mendeley should automatically merge the two.

If you have a set of papers where this consistently doesn’t happen, drop a line to support@mendeley.com and they’ll tweak the recognition algorithm.

Reply

7 Tom March 25, 2010 at 8:57 am

I’ve tried out Mendeley, and it looks very promising! One of my big issues with Papers was the lack of easy synchronization over multiple computers, and the ability to create shared collections, so this is a nice solution!

Jonathan: I looked into this a bit and it turns out someone had already requested this in the feedback forum (http://feedback.mendeley.com/forums/4941-mendeley-feedback/suggestions/275394-automatically-fetch-the-pdf-from-an-online-source). The issue seems to be that Mendeley would import the PDF directly without passing through your computer, so they could only import freely available PDF files with the ‘import to mendeley’ link. However, at the moment, it doesn’t even import the free PDF files on ADS, so I left a comment to that effect in the above feedback post. Feel free to log in to their forums and vote for this issue!

Reply

8 Gus March 25, 2010 at 12:43 pm

After reading the feedback link Tom provided, I’m unclear as to which of 2 scenarios Jonathan is hoping to see improved.

I first thought that Jonathan was referring to the Bibdesk like behavior(s) that (a) allow one to drag and drop a PDF onto a citation, adding the link and paper for storage and (b) looking for and flagging/repairing duplicates.

I really like feature (a) of Bibdesk, and no matter how I try, no field in a citation on Mendeley reacts to the drag n drop of a PDF. So one must use the manual “add” button do this, as Jonathan stated. This feature is the #5 most requested on Mendeley’s feedback page.

On to Bibdesk feature (b), Mendeley appears to try to duplicate check when one adds the PDF after a citation import via the bookmarklet. It flashes, “looking for duplicate” or something like that. PDFs dropped in twice will not twice import for me. Yet actually improved duplicate checking is the #1 most requested feature/improvement.

I tried as an example, “Mid-Infrared Extinction and Its Variation With Galactic Longitude” by Gao et al 2009 both from ADS and arxiv via the bookmarklet and from parsing the IOP PDF. Mendeley did not find or prevent duplication even though the ADS and PDF resulting entries have matching DOI! That should be a bright flashing light in my mind. Interestingly, Mendeley does not pick up the arxiv key from ADS, which I would personally like to see.

The second scenario is based on Tom’s feedback note and has to do with auto following PDF links on parsed web citations. When I went to the arxiv page for Gao et al, I found that the Mendeley bookmarklet does result in a linked PDF for the paper that I could view on either the Mendeley website or Desktop.

The Mendeley admin’s response on their feedback forum indicates that the issue of auto-following PDF links is not dealt with on a per site basis as Tom requested. You could argue that they should add the PDF linkages where ever one exists and let the user figure suffer the consequences if they cannot get to that PDF…

I would like to hear comments on the PDF citation parsing capabilities of Mendeley versus say Papers. The IOP ApJ Gao et al PDF from 2009 parsed perfectly, while another from 2008 ApJ article (DOI 10.1086/590364 — from UChicago originally I think) is a serious mess. I have my guess as to why but wonder if Papers would do better.

Also, if you use Chrome, there is an extension that works like the bookmarklet.

Reply

9 Jonathan Foster March 26, 2010 at 12:29 am

Hey Tom & Gus,

Thanks for looking into this in more depth and posting your comments. I understand why it is tricky for Mendeley to grab the PDF from ADS. That would be great, but I’m not holding my breath.

I was referring primarily to Gus’ scenario (b). I want improved duplicate checking or at the very least a way to manually tell Medeley — hey, these are duplicates, merge the entries (according to some preferences). That seems like it would provide an acceptable work-flow for me. Anyhow, this discussion would clearly be more productive on the Mendeley support forums…

Reply

10 Eli March 26, 2010 at 7:28 am

Does anyone know if Mendeley can combine author names? This is one aspect that is becoming a headache for me. In many cases I’ll have several papers with an author, let’s say “Simple, J”, “Simple, J. A.”, or “Simple, Joe A.” for example, and when I filter paper by author name I have to check out all of Simple’s variant names. Would be great if I could tell Mendeley that it’s all the variants are the same person.

Eli

Reply

11 Kelle March 26, 2010 at 5:11 pm

Apologies for the delayed posting of Mr. Gunn’s very useful comment; it was mistakenly marked as spam.

Reply

12 Mr. Gunn March 26, 2010 at 7:46 pm

Jonathan – Mendeley is working on supporting as many sites as they can, Support requests do help in the prioritizing thereof. As I said above, you should be able to combine duplicates in cases where you end up with a citation and a PDF entry for the same paper.

Eli – Mendeley has joined the Open Researcher and Contributor ID initiative, which is a community effort to establish an open, independent registry to resolve the systemic name ambiguity by means of assigning unique identifiers linkable to an individual’s research output. The researcher name issue is a complex one, but there is finally some progress in this area. We’re hopeful this will lead to a real and lasting solution to the problem.

Kelle – No worries 😉 Just trying to help.

Reply

13 Richard Scalzo March 29, 2010 at 12:16 pm

Hi all,

Eli — I think this is possible by dragging and dropping the duplicate author name onto the form you want to use in the “Filter By Authors” frame. It should then ask if you want to replace e.g. all instances of “Richard Scalzo” with “R. A. Scalzo” or whatever, and when you confirm, it will do a search-and-replace which is what I think you’re asking for.

I would definitely like to see improved metadata extraction from PDFs. For example, I find that the ApJ bibliography format is inscrutable to Mendeley and I get a lot of extractions where e.g. “Mnras” is the title (when titles aren’t usually given) — this is true even for recent IOP PDFs as well as astro-ph offerings. But the format is regular and if I could give Mendeley some kind of template for parsing them, just for a given journal, it ought to do much better. I don’t know whether this capability is already available (Gus, how did you get it to work?), or whether they’re planning on adding something like it. It’s probably not that crucial, just annoying — if they have that information available in the library, it ought to be correct without me manually entering every citation.

Modulo things like that, though, I really like the browser.

Reply

14 SciPlore / PDF Management May 10, 2010 at 2:06 pm

Hello,

if you like mendeley our software “sciplore mindmapping” might be interesting for you. sciplore mindmapping is a mind mapping application that allows you to integrate your pdfs and references with mind maps. the big advantage of this approach is that you do not only have a list of your PDFs but a mind map in which you can add additional information and arrange PDFs more flexible. And in case you really like Mendeley, you even can use Mendeley in addition to SciPlore MindMapping.

For a short demonstration of sciplore mindmapping see
http://www.youtube.com/v/jRHqLktIMWw&hl=en_GB&fs=1&rel=0&hd=1

To try the software (open source, Java): http://www.sciplore.org/software/sciplore_mindmapping/

and to read about how to write a phd thesis (or academic papers in general) with sciplore mindmapping see http://sciplore.org/blog/2010/03/02/how-to-write-a-phd-thesis/

Reply

15 Daniel McGravey November 27, 2010 at 3:56 pm

Mendeley is good but I feel like they track your activity similar to Pandora Radio and last.fm.

I like simple interfaces similar to itunes which is why I’ve stuck to using both PDF Stacks and Papers to manage my collection of medical articles and journals. They all do the job.

Reply

Leave a Comment

Previous post:

Next post: