Another MCN over. I’m probably going to have to wait a day or two for things to sink in for me to figure out what I’ve learned. Sometimes it isn’t even the presentations themselves, it is the casual mentions by a listener, or the casual conversation in a hall. My own archive, for instance, has recently wound down a project involving oral histories in New Orleans. We’re ramping up a project about WWII. Our oral historian will be delighted to continue the conversation I began with the person from the WWII museum in New Orleans.
I gave a talk, one of whose elements was about the ways we are starting to use cloud computing. To my surprise, we are relatively out on the cutting edge in this regard. It’s been about 15 years since I discovered that getting the IT department to take control of my servers meant that they had to wear beepers, not me. Some time after that it sank in that for most businesses, there is no reason to host web servers locally, and many good reasons to host them on slices, or actual colo servers, maintained in faraway cities by people who worry about connectivity 24×7 for a living. The fact that our services now aren’t on a box in a cage to which we could, should we really want, travel to and touch doesn’t seem like a particularly large transition. Quite the contrary–the idea that I can add and provision additional servers (or ditch servers I no longer need) in minutes seems like an appealing extension of the idea. It’s like making the transition from having to wire every computer in the office to a specific tether connected to a specific node back on our network panel, to simply using wireless and being able to think of capacity in a simpler, more liquid fashion.

For our specific application, backing up to storage in a computer cloud run by Amazon, the magical moment came when I first discovered NetApps boxes and realized that I could separate data storage from the servers that had to use the data, forever ending the need to copy and maintain duplicate sets of data–often duplicate sets of much data–on many identical servers. In return for sharing our experiences, however, several people have mentioned the joint Fedora/DSpace project to create a cloud specifically for institutions like my own, DuraSpace (“…Space,” the new Mellon Foundation frontier…. Oops.). We’re ready to volunteer and see if this makes more sense than using a commercial provider like aws.amazon.com.

I was gratified to realize how many Drupal users there are in MCN. This open source content management system is being used to power a growing number of our website services. One of this year’s challenges will be to find ways to collaborate so that we can share specifics, and share brainstorming and problem-solving. I think I’m going to look very good back at home when I show of the dashboard module developed by Rob Stein’s team at the Indianapolis Museum of Art. If I can talk some of these MCN Drupal users to collaborate on interfaces to the Fedora Commons repository, we’ll be able to make our software development budget go much farther, much faster, and likely help mature some still-simple tools in ways that will solve problems for all of us.

One of the most exciting sessions was reviewing early Semantic Web work by an amazing team from the Met. For those new to the concept, the idea underpinning the Semantic web is to build enough smarts into our web pages (some of which will get there using automated text analysis tools) such that one can ask better questions of data. Instead of “searching” for a match with “Emma Lazarus” on a website, one day I will be able to get the same results by asking about “the poet who wrote the poem on the statue of liberty” or “Jewish poets of the 19th century” because the implicit data from various snippets about Emma Lazarus, will have enabled the underlying semantic web engine to put bits of information together and answer increasingly sophisticated questions about what data are actually present on a site. This is big. One of the references? The accurately named (and delightfully written) Semantic Web for the Working Ontologist: Effective Modeling in RDFS and OWL.

The final sessions I attended contrasted two different approaches to Open Source software planning and management, whether intentionally or otherwise–that of Archivists Toolkit and the still-largely-on-the-drawing-board CollectionSpace (nee OpenCollection, now picking up on the new “…Space” meme). The two projects address similar issues, but listening to them described it is clear that merely saying “Open Source” doesn’t tell you much about the viability or quality of a given software project, any more than “commercial” does. The label only describes part of the development and maintenance process. As you consider whether to deploy either open source or commercial software, your major costs lie in the software customization and long-term maintenance. Software license fees (or lack thereof) are a very small part of those costs.

Instead, the critical issues surrounding software viability really lie with the community of vendors and users (along with your clarity as to your needs, and the actual fitness of the software to meet those needs). That, of course, is where an organization such as MCN excels–in helping you find that community, share experiences and problem-solving with peers around the world. Having said that, it will be most interesting to see how the MCN Board shapes changes to MCN community tools–currently this blog (co-run with AAM’s Media and Technology committee), the MCN website, and the MCN mailing list–adding new tools, or changing how these are run–to take advantage of current web affordances for building and sustaining community.

See ya’ll online.

P.S. Late addendum. David Dwiggins took some incredible notes of several conference sessions. Take a look at Dave’s Conference Musings. Enjoy.