I’m not seeing much blogging about MCN2008. (I suspect because, as usual, too many of us are having too good a time networking and attending sessions, abetted by the rather unfortunate lack of technical facilities–wireless, comes to mind–that would often be present at a gathering of, uh, techies. Expect some changes as we prepare for next year.)

Let me try to fill in one gap, at least, by posting the slides from my demo yesterday, “Open Source Digital Asset Management/Preservation on the Cheap.” In short, yet another forceful, masterful variation on my usual theme: small institutions have few resources and need to maximize use of the inexpensive, cutting edge, and especially open source technologies to have the impact, and to engage our audiences. If we do this well, those big institutions will try to learn from us .

In this particular case I have been obsessing about how to reasonably assure the long-term preservation and integrity of about 6TB of digital assets at the Jewish Women’s Archive [jwa.org]. Our mission is to uncover, document, and disseminate the stories of North American Jewish women. We exist on the web, only, so if we lose that digitally documented history, there is usually no paper to fall back on. We’re it.

Last year, we took an aging collecting of audio artifacts–primarily cassette tapes and minidisk recordings–and got them all transfered to accessible digital forms. (In the case of minidisk recordings, this means copying recordings from Sony’s proprietary digital minidisk format into common WAV files). These are all stored internally on a very expensive (to us) 6TB RAID server in the air conditioned room where we keep our network gear and servers. RAID is a wonderful way of combining lots of disks with logic akin to a Sodoku puzzle. If one disk dies (and sooner or later, all hard disks die), we can slip in a new disk, and the computer logic will solve the Sodoku puzzle, filling in the missing information, and all is recovered.

But, RAID doesn’t save us from a local catastrophe. What if there is a fire? What if some unforeseen natural disaster strikes Boston? What if thieves break into our office and steal our servers?

In an ideal world, we would team up with other organizations, and each back the other up. As a practical matter, I have no budget to back up anyone else’s data, and the same, alas, is true of our peers. I had to look elsewhere.

Amazon.com offers some amazing, low-cost, internet-based infrastructure services called “Amazon Web Services” [aws.amazon.com] (AWS). It costs us 15 cents/month/gigabyte to store information on their “cloud computer” storage servers. (The service is called “S3″.) We purchased a $20 utility, JungleDisk [www.jungledisk.com] and can “drag and drop” our files to S3. We then took things a step further. We already using version control software, Subversion [subversion.tigris.org] to enable us to keep track of our software and to roll back changes as needed. What if we used Subversion for documents that change–the XML metadata files, transcripts, etc. Then, we wouldn’t just have the most recent versions of things, but would have a history of the authoritative version of each document. No more “FINAL REALLY FINAL transcript.doc” files to compare to “2008-03 FINAL transcript.doc” files. Subversion, of course, is open source. That means that instead of us being dependent on a commercial vendor developing new features every so often, we can rely on our fellow users (ourselves, included) to fix bugs and add new features as they matter and as we have the time and expertise. Quite a win-win choice.

As it happens, the instant we introduced a modicum of version control, our Board and upper management were carried away by the synergies of having a “real” repository such that all of our assets could be online and accessible. The not-quite-finished-in-time-for-this-conference next step involves use of the open source Fedora Commons [www.fedora.info] repository framework in a minimal “good enough for our administrator to cut processing time by an order of magnitude” setup, to be followed, next year, by a host of tools to integrate our open source content management system, Drupal (www.drupal.org) with Fedora for even greater synergies.

Later in the day I spoke at a panel on hosted digital asset management systems and their desirability. One listener complained that there were no open source solutions to this pressing issue, and no standards. I maintained that this was coming quickly, exactly on these lines (Fedora+Drupal) in concert with METS and subsets of same including PREMIS, Dublin Core, and a host of Cultural Heritage Institute-specific ontologies and standards. It’s time to put them together.