Mendeley and social citation software

Expanded repost of a comment I made on a post at the University of Windsor (I think) Leddy by Design blog.

UXLibrarian first muses about the difference between ‘weak’ social software that doesn’t really create or engage much of a ‘network effect’, vs ‘strong’ social software that does.

I would like to see students and faculty be able to take links to the research material from our online library and use and re-use these links within in their own readings lists, blogs, and email conversations.

So I am hoping that every book or journal mentioned on our website is expressed in such a way that a tool like Zotero is able to render it into a re-usable link.

UXLibrarian considers using COinS for this purpose, but then realizes this is “trickier than first thought”, for a variety of reasons. I can think of more reasons that make this tricky, including that, you know, nothing actually uses COinS much but Zotero.  COinS really have a variety of problems even if more software did use it (or perhaps reasons that more software doesn’t), not being URLs, they aren’t really “of the web”, they are just tricky to deal with.

But the whole problem here is that a citation to a book or scholarly article doesn’t really have a URL. Such things may not be on the web at all, or may be on the web at several places (in the case of scholarly articles) or even hundreds (a book listed, probably without fulltext but with summary citation info, in hundreds or thousands of places). We’re not dealing with web pages, we’re dealing with scholarly or creative or informational works that exist in many cases primarily off the web, and in some cases in multiple places on the web where different people are licensed to see different places. What a mess! And that’s what complicates UXLibrarian’s final musing: “Addendum: would something like TinyURL or other link shortener be a viable option?”

The trick is, what do you want to link to, in the end?  One particular version of a scholarly article, that some people are licensed to see or others are not?  Might be good enough for your users, until you change your providers. A generic citation that will somehow identify the link and get the user to their appropriate licensed copy?

That’s what OpenURL is for, and COinS in particular is an OpenURL un-moored from a particular libraries “link resolver” (the thing that gets you to YOUR appropriate copy).   But now we’re back at COinS, which are a pain to deal with, and have haphazard support.

You could generate a complete OpenURL identifying the citation and tied to your local link resolver, and make a tinyurl out of it — but you’d still have to generate the OpenURL somehow (which is basically no harder and no easier than generating a COinS) — and now it would still be tied to your particular institution, not really something that does someone from another institution much good, and still not the kind of URL that we’d like to serve as a stand-in identifier across the web for what we’re talking about.

It’s all kind of a mess. is attempting to be “strong social software” (in your terminology) for scholarly citations. Not sure how well they are or will succeed, although they seem to be gradually succeeding.  You could use a simple link to the mendeley page for an article (if it’s in mendeley, or import it into mendeley if it’s not).  But your users still can’t get to their licensed copy from your institution — ah, but I have been told by mendeley folks that mendeley will be adding outgoing OpenURL links, where a user can choose in preferences their institutional affiliation (like Google Scholar does).

If Mendeley can pull all this off well, they will definitely be filling a valuable niche — and maybe you could indeed just link to the Mendeley page for an item. I don’t know how many books (as opposed to articles) let alone say movies or musical albums that are in Mendeley though, or what would be the quickest way to add such an item to Mendeley if it’s not.

(Mendeley does have an API; I can imagine integrated editing software that lets you enter a citation , searches Mendeley for it and embds the link if found; if not found adds it to Mendeley and embeds the link to the thing added.)

Mendeley seems to be kind of predisposed to dealing in URLs, and a book doesn’t really have a URL (That’s the whole problem in the first place here!) — this same predisposition seems to give Mendeley problems sometimes with having multiple records for the same citation not realizing it’s the same citation, because the article is available at multiple URLs.

Mendeley is also focusing on “papers”, not books or videos or what have you –but that doesn’t mean other stuff doesn’t sneak in, people cite things other than ‘papers’. Here’s a record for a video in Mendeley, although you couldn’t actually easily tell it’s a video from the Mendeley record. (I am familiar with the work, plus the clue “Publisher: Zeitgeist Films”.

But see there, I sort of kind of just linked to a video in a way that users can “take links to the research material from our online library and use and re-use these links within in their own readings lists, blogs, and email conversations.”

But the moral of this story is : the information domain we work in “bibliographic” or “scholarly” or “informational” citations, is a really complicated one!  Which is what makes it interesting for us library hackers, but also why it is not surprising we have such trouble making it all neat and clean.

5 thoughts on “Mendeley and social citation software

  1. I hear your point about COinS. I would have thought they would have seen better support outside of Zotero and Mendeley.

    I’m glad Mendeley is proving useful in this respect and other kinds of scholarly material besides just papers are certainly welcome!

  2. I think I was unclear in my post.

    For books and journal articles, I am still very interested in providing both a link to our native resolver *and* a COinS link. I’m trying to work towards a scenario where the user is able to add an article, book, or journal title from the library website into Zotero (using COinS), which they then can drag and drop into their workflow.

    Our course reserves are COinS enabled, as is our catalogue, and very soon our SFX menu will have COinS too. It makes sense for our website to be next especially as Zotero’s standalone client is in alpha release.

    What’s unclear to me is how we should to make ‘indexes’ more like citations.

    It should be easy for a Psychology prof to add a link to PsycINFO to a course page without having to worry about a proxy prefix.

    Our users might be more inclined to re-use the links to our databases if the links were shorter and had meaningful names and wouldn’t break overtime. My first inclination was to use OCLC’s PURL service or set up a server but that seemed like overkill for something so simple. I know that UofT runs their own URL shortening service from their SFX menu, so my addendum was really a start of a whole new train of thought ;)

  3. Yeah, I didn’t realize you were talking about databases/indexes/ platforms type stuff in your post. Links to web-hosted databases…. here’s how I handle that where I am. We run Xerxes as a front-end on top of Metalib. We also use Xerxes as our “database of databases”, our user-facing directory of databases. One thing Xerxes does for you is give you that kind of shortcut redirect URL, which will send the user through the proxy if the resource is marked as requiring it, etc.

    For instance, here’s the “directory” page for FirstSearch WorldCat in our Xerxes:

    The “Go to this database!” link on that page is actually:

    That can be bookmarked, and what accessing that URL will do is actually redirect you to FirstSearch WorldCat, through our EZProxy. Don’t be fooled by the “/proxy” in the URL, it only sends this database through EZProxy because the Metalib resource is marked as requiring proxy. Here’s another one for a free resource that does not send you through our EZProxy:

    So these /datbases/proxy/$ID links will keep working if we change our mind about whether a resource should be proxied. It will keep working if the actual URL to the database changes (perhaps the vendor has made a platform change, with a new url), so long as we update the Metalib record. It will keep working no matter what — as long as keep using the same Metalib IRD id (that’s what that “JHU02546” is) for the same database. Sadly, the most convenient Metalib knowledge base management processes don’t always do this, which is a pain.

    So using Metalib as the basis for this is not neccesarily ideal, and you wouldn’t use Xerxes _just_ for this; but if you’re already having to keep track of things in Metalib and don’t want yet another store of connection details to maintain, and you’re already using Xerxes because it provides a better interface to Metalib…. then this is good value added in Xerxes to use it as your ‘database of databases’ providing this pass-through redirect feature. If you’re not using Metalib/Xerxes… well, maybe the pattern is instructive for other ideas hopefully based on existing stores of database connection info you already have, because nobody needs yet another one.

  4. I recently came across this article by Tim Berners-Lee on the concept of resource in “Uniform Resource Identifier”:

    So URIs a) identify things (accessing them is a separate matter), and b) don’t just refer to web pages, but can also refer to more abstract things, including books.

    Too tired ATM to draw out the conclusion :-)

  5. The problem maybe is that there are too MANY potential URI’s representing, claiming to represent, or used by others to represent, a given printed book.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s