Saturday, April 30, 2005

1 + 1 > 2--or Gather locally, share globally

Two developing web-based systems have a potential to produce an incredible synergy.
  1. Social bookmarking for academics:
    • CiteULike: Richard Cameron designed an built it in November 2004 and has run privately since then.
    • Connotea: from Nature Publishing
    • biologging: Alf Eaton's community website for biomedical researchers. This one has other features but works best with HubMed (it's a really cool alternative interface to NLM's PubMed
    This kind of system offers an amazing arena for scholars to share and develop language for describing and locating work in their own areas of study. [I'll spare you the long winded description of theoretical models of contextualization as a social process in the construction of meaning. Trust me, this is amazing.]

    A major limitation of these services is that they do not easily capture bibliographic information about the references collected (author, title, sources, date, etc.). All of this information is invaluable for the more sophisticated search needs of academics. I might want, for example, to view only those articles written before 1999. Also, getting a bunch of links to articles on a web page is only half the battle. Eventually I'll want to use the citations in a paper. I'll need that bibliographic information.

    Also, if I attempt to share citation links online I run smack into the old 'appropriate copy' problem. Maybe I point to a version of the article that's on the publisher's web site but your only access is via Academic Search Premier.

  2. Latent OpenURLs
    OpenURLs can, in theory solve both problems. They can capture that bibliographic data in a rule-governed format that makes harvesting bibliographic information a relatively straightforward process. And they include an element that refers to a link resolver.

    Unfortunately, as they are most often generated now, OpenURLs bind the bibliographic information with identification of a specific link resolving systems that cannot, by definition, be relevant or useful to all readers. That is, to use OpenURLs I have to find a way to make them point to the right link resolver for different readers.

    Latent OpenURLs come to the rescue. They provide a means for embedding metadata via OpenURL specs in regular HTML code without specifying the particular link resolving system to be called upon. Instead a reader's browser can, for example, use a bookmarklet or browser extension with a very simple java script to call the appropriate link resolver from a Latent OpenURL. (This is misleadingly simple. See Daniel Chudnov et al's article on this in issue 43 issue of Ariadne (April 2005) for a full discussion.)

So the social bookmarking manages intellectual access to the resources while Latent OpenURLs provide seamless/transparent linkage to the appropriate version of the resource cited.

No comments: