I spent the better part of today playing around with bibliographic management tools. At the beginning of the day, I had several (small) EndNote libraries that I’ve been adding to for my PhD as well as a few other projects. I am, however, rather lazy, and I realised this morning that getting things I find on the web into EndNote X3 was far too cumbersome for my liking.
Sure, I could have walked up to campus in the torrential rain (and I’m not exaggerating about the rain, either) to get a copy of X4, which can extract metadata from PDFs. But, you know… lazy (and far too much of a princess to get that wet).
Having watched Howard Rheinghold’s explanation of how he uses “Twitter, search, Diigo, Delicious, DEVONthink, Scrivener to find, refine, organize information –>knowledge“, I started to think that maybe I should be looking for a new tool. While I was pondering downloading a trial of DEVONthink, Kathryn Greenhill tweeted the suggestion that a combination of Mendeley and Zotero, with full text stored in Dropbox, might work. I’m a Zotero fan, but I’ve been worried about scalability and the logistics of using it to manage something as big as a PhD. But essentially, I was already sold on Zotero as a tool. Mendeley, however, I had no experience with, so I decided to download it and have a play.
It was love at first site.
But back a couple of steps… The path to get my data into Mendeley was fairly smooth… I exported my EndNote libraries and then imported them, one-by-one, into Zotero. Then I did a bit of organisation to put my references into collections. After that, I downloaded Mendeley, installed it, signed up for an account, and turned on the function to continuously sync with Zotero. And bingo! All my references magically appeared in Mendeley.
Here’s what made me fall instantly in love with Mendeley:
- The idea of being able to extract metadata from PDF files to populate item information. I say ‘idea’ because in reality, the examples I tried were great big failures, resulting in references that looked nothing like what they should have – wrong author, wrong title, wrong date, wrong publication, wrong topic… My intention had been to then export references created this way to Zotero – another great idea that didn’t work in practice.
- Mendeley prompts you – with an unmissable dialogue box at the top of the data pane – to check and verify references when you import them. Given my experience with the quality of data imported into EndNote, I was really excited to see that new references were flagged to check, and just as importantly for lazy me, that with a single click I could search Google Scholar by title to verify the details. The final win on this front is that any details that aren’t correct automatically get fixed, and voila! You have beautiful data. If the item can’t be found by a title search, you can add the DOI for the item to the DOI field and retrieve the data that way – again, with a single click.
- I really, really like being able to see at a glance what I’ve read and what I haven’t. Mendeley allows you to toggle a small dot beside each title from grey (unread) to green (read). Yeah, I could tag stuff as unread in Zotero, but I know from using this tactic in delicious that I’ll never remember to go back and edit the tags.
- Functionality to view and annotate full text attachments is available in Mendeley – and it’s nice functionality, too. For a blissful couple of hours, I saw myself forsaking my nemesis, the office photocopier, forever.
- Zotero integration means I can continuously download references from Zotero to Mendeley without any effort whatsoever. This appealed to me primarily because I wanted to use Zotero to capture data while trawling the web, and to integrate with Word when I’m writing, but I wanted to use Mendeley to verify data, and view and annotate full text.
Here’s where it all fell to pieces:
- Syncing only works in one direction: Zotero to Mendeley. There’s no functionality in either Zotero or Mendeley to support automatic updates of Zotero. Strike one.
- I kept getting duplicates in my Mendeley library. I puzzled over this for ages, until I realised it was only happening for references that came from Zotero, but which I had edited in Mendeley (to correct errors in the data). Turns out the automatic syncing functionality kept pulling in the references from Zotero, because they didn’t look the same as the references in Mendeley. Strike two. Which leads me to my next point…
- Deduping is tedious. (That’s such a library geek word, isn’t it?!) Identifying and removing duplicate records in Mendeley is an entirely manual process. Find, compare, right click, delete. Repeat. And then repeat again because you forgot to turn off the automatic sync with Zotero. And then repeat again because you turned it back on just to double check your hypothesis. Strike three.
So, my short-lived love affair with Mendeley is over. I’ll watch developments, though, because I’m definitely open to giving it another try if and when an an automatic function for identifying and removing duplicates is added, and Zotero sync becomes a two way street.
The task monster in me feels as though I wasted a day messing about… But as one wise tweep of mind said to me today: I know that part of working with new tools is spending some serious time messing around.
My next challenge will be to have a play with Scrivener and decide if it will work as a repository for my notes, and potentially a tool for writing my whole thesis (and all the other smaller outputs along the way). Part of that task will be working out how Evernote (which I’m using now for note taking and brain storming) and Scrivener will differ in terms of my workflows (aside from the fact that Scrivener is better suited to writing content). As a part time student, I’m going to be working on this beast intensively for periods, and taking chunks of break time while I prioritise teaching and other research commitments. I’ll need to be really organised to manage that. Plus, I want to be able to store content I produce along the way in a manner that makes it easily reusable in different forums – from my proposal, to my confirmation of candidature materials, to journal articles, to the thesis itself. But that is most definitely a challenge for another day.
Zotero, too, can retrieve data from pdfs:
http://www.zotero.org/support/retrieve_pdf_metadata
I wouldn’t worry about scalability – Zotero is getting faster and more robust with every update.
Ok I see now why you directed me to your blogpost
Just watched the zotero vid….looks pretty awesome.
But now I have a new toy to procrastinate with–Scrivener :-\
Hi Kate! To speak on behalf of Mendeley, we really wish the sync could be two-way too, but unfortunately, they don’t have an API that we can write to. You put that down as a strike against us, but the M->Z part of the sync is really up to them. The issues you had with duplication seem to have arisen from that. We’re working on a better way to detect these kinds of duplicates on our end, but until then I hope you won’t hold it too much against us. Until their API is ready, you can import your changes in Mendeley back to Zotero by exporting from Mendeley and reimporting the export file into Zotero.
It strikes me that you could take advantage of all the things you love about Mendeley as long as you leave the sync off. If it helps, Mendeley has a bookmarklet that you can use to import references directly into your library.
Thanks for giving Mendeley a try and blogging about your experience. Hopefully Zotero will get their API finished soon and we can get real two-way sync going.
Extraordinary! So pleased that you spent time on this as I am a novice in bibliographic management tools so your post really helped me! I can’t wait to get myself better organized- though I think it will be a big journey for me, your post has really guaranteed that I will get it done. Let my experimenting begin.
Thanks for sharing your experience. I, too, tried and abandoned Mendeley while sticking with Zotero.
Using a consistent file structure on my computers and SugarSync or Dropbox allows me to keep my PDFs accessible and backed up. And Zotero is just wonderful as a total reference manager for all types of cites and notes, tags, etc.
I use a very consistent naming convention for all my PDFs. Store them in a folder that is identical on all my computers, and use SugarSync to keep the machines in perfect alignment. Been doing this for over a year now. Have nearly 2K refs. And have a rock-solid arrangement. I can go to any of my machines, open up Zotero, and access a PDF (even one modified by iAnnotate on the iPad).
Works great. (Like you, I previously experienced a Mendeley mess of multiple duplicates, renamed files, copies of PDFs in multiple folders, and on and on. Just didn’t perceive any upside and lots of downside.)
If only Zotero integrated with my beloved DevonThink Pro (and had an iPad solution)!
BTW, I am a Scrivener user as well. Wonderful writing environment. Superb for “chunking” content, rearranging, and having it automatically assembling itself into a cohesive whole. I’ve put together templates for both research papers and output styles suited to my final draft (which I do in Pages which is then output to Word format). A new super-duper version of Scriv is due later this month…
Appreciate your detailed story.
Continued productivity!
All the Best,
Don
PhD student (proposal stage)
Human Development
thanks for your comment – i’ll definitely be giving mendeley another go if and when two way sync gets going.
hey sebastian. i did know about zotero and retrieving metadata from pdfs, but i’d also read that mendeley did this better, and the first few pdfs i tried in zotero didn’t work well. but i’ve since used it quite a bit, and it works just fine. as any good librarian will tell you, crap data in, crap data out – it’s all about the quality of the metadata in the pdf.
yeah, do want scrivener. but not sure i can justify losing the time i would spend playing with it… oh well… maybe over christmas?
i highly recommend using a bib management tool to organise your course readings for your teaching, judy. i use folders for each unit and a tag for the week number. that way, i can export a bibliography for my slides/notes (students find it amusing when i don’t cite properly, so i’ve learned to include a full bib), and when i teach a unit again, i’ve got the readings from last time at my fingertips. while i often don’t reuse the readings (preferring to use current material), having them there reminds me why i chose them, which helps me decide whether to reuse, and helps me to select new ones.
thanks for your comment, don. now that i’m starting to build up a stash of pdfs in zotero, i’m rethinking keeping them there, and i think a common file structure across computers and dropbox might be a good option.
i think i’m going to have to get scrivener… i’m a big fan of writing in chunks, so i proliferate word docs, and i think scrivener would help me keep everything sorted. i might wait and see what this new version looks like.
would love to here more about how you integrate devonthink pro into your workflow.
Kate – Mendeley is more innovative in how it gets pdf metadata, but doesn’t actually appear to produce better results
http://musingsaboutlibrarianship.blogspot.com/2010/07/extracting-metadata-from-pdfs-comparing.html
Zotero doesn’t actually extract the data in the file. It first looks if it can find a DOI – and looks that up on Crossref. If that doesn’t work, it uses the first result from google scholar – for the first couple of sentences or so I believe. In general that works quite well, especially because google scholar is also improving.
Sebastian – I’m familiar with Aaron’s post and spoke to him a little bit about it. He didn’t do the test of Mendeley while logged into his medneley account, which meant that only the information that could be scraped from the document without any online lookups was used. As the Zotero guys will note, this really isn’t a good approach, especially for older papers.
Mr. Gunn – I know. That’s why he amended the blogpost – in red, so it’s not like it’s easy to overlook – after re-trying while he was logged in. What he found was that Mendeley did much better than originally and about the same – certainly not better – than Zotero. Which is exactly what I wrote.
[...] is worth a look for a discussion of some of the challenges of choosing and using multiple options. http://virtuallyalibrarian.com/2…Insert a dynamic date here BIU @ @ ReferenceEdit [...]