Still, the accomplishment was one we paid forward a while back when we made a couple of decisions about design and architecture for the digital collections.  The most important was the decision to eschew “black box” vendor solutions in favor of developing our own system using open-source and freely-available software tools.  That choice gave us control over our metadata and content, which, as I mentioned, set up the use of cooliris for search result sets, which in turn gave us RSS feeds to tweak and re-use.

These days, we’re thinking about how to extend or re-develop the “Tripod” system that serves as the front end for our digital collections, especially in light of ongoing progress that we’ve made with the back end, a repository and metadata editor, a.k.a., the “Trident” project.  The experiences with cooliris and DukeMobile have reinforced in my own brain a couple of ideas that I hope we’ll embrace as we go through this process, namely, syndication and aggregation.  I know I’m not breaking any new ground here, but I wonder if these concepts have gained real traction in the world of library digital collections.

Syndication involves opening up pipes of content.  To some extent we can already do this, as we did with the RSS feeds for cooliris and DukeMobile, but the API that we have for it is somewhat accidental.  I’d like to reconceptualize that API and make it more central to the architecture.  We also need more control and flexibility with respect to the kinds of lists we can create, whether through metadata elements or the assignment of relationships between objects.

The objective of syndication is to enable users with a diverse array of client programs to pull selected content into workspaces that are meaningful to them.  The pattern is one of distribution or diffusion of content, but the best way for a manager of collections to cultivate that pattern is to achieve a centralized control over that content.  Then a single, unified application framework can serve as a mechanism for opening up those pipes.

Aggregation is the consolidation of content from disparate sources.  It seems like the opposite pattern from syndication, but it’s really the same “pipe” seen from the other direction.  I think this idea has some deep implications for the appeal of digital collections.  Duke’s collections are “here,” ECU’s “there,” NCSU’s “over yonder,” and so forth and so on.  From a researcher’s point of view, the collections might offer more gravitational pull if they were all accessible via a single discovery interface.  I think consortium-level aggregation of digital collections is an idea worth exploring as a policy.

We know it’s possible as technology.  Consider the OAI-ORE specification, which conceptualizes a digital object as a “resource map”.  One mode of implementing the resource map is to use the Atom syndication format.  So OAI-ORE gives us a way to describe an item in a digital collection that immediately lends itself to syndication.  In turn, an application that implements the “other” Atom — the publishing protocol — could aggregate digital collections content within a single interface.

Aggregation and syndication are really two sides of the same idea, or complementary ideas, one that follows from the other.  In general, I’d like to see us in the library world spend less time building interfaces for digital collections, and concentrate more on producing content, describing it in meaningful ways, mashing it up, building connections within and across collections, writing blog posts about it, producing videos about it, writing songs and hard-hitting melodramas about it, and bragging about it on social networking sites.  These activities will lead to good publicity, which causes the release of endorphins, in a repeating cycle of good.