It's also important to stress that this is all happening within the intellectual context of ISAW. In other words, my new colleagues have pushed on these issues in interesting ways and I can take advantage of their previous work.
As in, let's talk about Zotero and the role it can play in providing a sustainable bibliographic framework for digital scholarship. This is already happening at the ISAW-hosted Pleiades project, and that lets me take a very practical approach to writing about the issue.
The Pleiades Zotero library is at http://www.zotero.org/groups/pleiades/items. It includes the item http://www.zotero.org/groups/pleiades/items/72922905, which is the Zotero record for the article
Coastal Sites of Northeast Africa: The Case Against Bronze Age Ports. In case it isn't clear, the point of the article is pretty much that there weren't any.
In yesterday's post, I talked about citing secondary scholarship. Today, I'm interested in the mediating those citations through Zotero bibliographic records.
The same basic pattern would apply:
'<a rel="dc:references" href="http://www.zotero.org/groups/pleiades/items/72922905">White and White 1996</a>'is a reference to the work described in that Zotero record. I am interested in the extent to which it is necessary to indicate that it is not a reference to the Zotero record itself but let me put that off for now. More relevant here is why use Zotero to establish unique identities for cited works at all?
The most compelling reason is that not all works will have such identifiers and Zotero allows one to create these. For example, http://www.zotero.org/groups/pleiades/items/72922931 is the Zotero record for an article that has no direct representation in Worldcat and which isn't online (I don't think). I.e., you're on your own in terms of a stable URI for this title.
And since consistency is good, it might be appropriate to create Zotero records for all cited works in a digital publication and only point to those.
This approach is also attractive because it allows linking to digital representations of titles as they become available. For example, the record for Coastal Sites could be linked to http://www.jstor.org/stable/40000602, which is the JSTOR record. A more compelling example is this link: http://hdl.handle.net/10.2972/hesp.79.1.1. That will take you to the Atypon-Link version of J. Cherry and W. Parkinson's Hesperia 2010 article on lithics from SW Greece. As the volumes of Hesperia role over into JSTOR once they are past the 3 year wall, the Atypon URI will be either matched or replaced by an equivalent JSTOR URI. A Zotero record can have links to both versions without requiring any updating of the digital publication which points to that record.
And if more than one digital publication points to a Zotero record, that equivalency should be discoverable. I like that.
The big potential downer is: do we trust Zotero to be around for the foreseeable future? Or at least, will these URI's work over the very long term? I don't know the answer to that. This is one reason to ensure that each digital publication "knows" bibliographic metadata for all the citations it makes. Centuries from now, that information may be useful in tracking down readable versions of titles.
And here'a a finishing twist. Regardless of which tool is used to generate URI-based unique identifiers for cited works, that same tool could (should? must?) be used to provide URI-based unique identifiers for the digital publication itself.