Berlin, Germany Last month HatCheck spoke with Matt Zumwalt, MediaShelf—a company that works with businesses and organizations to create digital asset management solutions with Fedora—as he was settling into a flat in Berlin. Currently Matt is on what he calls a worldwide “Listening Tour.”
MZ: MZ: The Listening Tour is going really well. For my first formal interview, I spoke with Ari Davidow at the Jewish Women’s Archive on the last day of 2007. He’s actually the only librarian/archivist type who I’ve interviewed so far. I’ve been trying to talk directly with people who create content, since ultimately that is who a content repository needs to serve.
In addition to Ari, I’ve sat down with a Tibetan Lama in upstate New York and a photographer/performance artist in Minnesota. I’ve also got a couple of interviewees lined up here in Berlin, including an astronomer, an ethnographer, and the booking manager for a popular jazz session cafe. So far, the interviews have been really wonderful. I’ve already encountered things that forced me to reconsider a lot of assumptions about repositories.
I had originally planned to make little radio-style edits of the interviews and put them into a Listening Tour Podcast. That quickly got nixed when I discovered that creating a good podcast could easily become a full time job! I still hope to do that [podcast], but not for at least a few months. In the meantime, I’m writing up accounts for the MediaShelf Blog and accumulating the recordings in a home-grown Fedora repository.
I’m actually having a lot of fun archiving the interviews because the repository I’m creating is a little bit free-form and wild. It’s safe for me to try out new things and to think like an end-user instead of a repository architect. I can get away with just dumping objects in there with all sorts of haphazard metadata because the repository is mine to play with. This is actually where Fedora shines, because I’m able to make this messy repository now, and clean it up later. It’s all versioned, and it’s completely flexible. I’m learning a lot from this free-form tinkering, and will learn even more from cleaning up after the fact.
HatCheck: Another interesting project is the European Digital Library Project (EDL). The national libraries of many European nations including Belgium, Greece, Iceland, Ireland, Liechtenstein, Luxembourg, Norway, Spain and Sweden have come together to create a consolidated search portal. You attended a conference in Frankfurt that reported that initial content has been digitized. Is the search and browse experience geared towards ordinary European citizens?
MZ: With 200 digital archivists and librarians speaking candidly about their work, the EDL conference was a perfect stop on the listening tour. At the conference they unveiled a maquette of the portal’s user interface. You could see the influence of sites like YouTube, which makes sense because they’re trying to target everyday users. You could also see components that try to take advantage of the depth of formal metadata the EDL has at its disposal. I think the work they’re doing is really exciting. Europeana: Connecting Cultural Heritage will go online later this year.
Hatcheck: What happened at the workshop you led at The Bloomsbury Colleges consortiumlast month in London? As I understand it you took a roomful of Fedora users through a variety of hands-on topics such as setting up Fedora on Amazon Web Services (aka. Fedorazon), understanding the new RESTful API (which MediaShelf contributed to Fedora 3.0), and experimenting with Fedora’s new Content Model Architecture.
MZ: The whole Fedorazon thing is really interesting. Utility computing is definitely a trend that’s here to stay, and Amazon has taken a bold step into that arena. Amazon’s Web Services are still very new, and very raw though. Right now when you set up an EC2 instance, it’s like you just took an old computer out of your closet and installed Linux on it. You don’t have any of the modern tools and conveniences for managing servers. Despite this, once we had created EC2 images with Fedora pre-installed, I found myself using them over and over. Definitely watch this space.
The REST API in Fedora 3.0 Beta is exciting because it allowed me to put Fedora in front of a roomful of people and give them a real, direct sense of how you can use it.
At the end of the day, we played around with the new Content Model Architecture (CMA). One group at the workshop tried to “push” the examples, making them do new things. We expected our first attempt at rewiring everything would fail, but it actually worked. This was pretty encouraging.
HatCheck: What are you hearing with regard to what people would like from Fedora both in terms of functionality and additional core services?
MZ: Two years ago, when I started working with Fedora it was still pretty raw. I started MediaShelf based on the conviction that Fedora should be easier to use. I put together a team of talented friends and set out to achieve that. Meanwhile the Fedora core team have done a huge amount of really great work. Compared to two years ago, Fedora is much more flexible and much easier to use, but there is still a long way to go. There are some things, like workflow, that I’m not so sure Fedora should be responsible for. Other things like messaging, transactions, and clustering are really important and definitely should be supported by future versions of Fedora. There is also the ongoing challenge of getting the authentication and authorization stuff to sing. There’s been a lot of talk about that on the Fedora-dev mailing list lately.
Also it would be really cool to make Fedora accessible through a JSR-170 interface. This would be a big advantage for Fedora adopters.
HatCheck: How is 3.0 beta with the new Content Model Architecture working?
MZ: It’s really exciting to see the CMA taking shape. I see Fedora as an architecture before I think of it as software. For me, Fedora’s primary purpose is to provide a really solid conceptual framework for making sense of the complex problems around digital content repositories. The CMA carries that framework forward into exciting, uncharted territories.
HatCheck: Where do you see the biggest FC growth—what type of repository, eResearch, Library/archive, ePublishing etc.?
MZ: I’m not sure that I would say any of these areas is more active than the others. We’ve worked with people in all three of those fields this year, and we’re constantly hearing of new projects all across the board.
HatCheck: What kind of additional materials or documentation would help you work with folks better or more efficiently?
MZ: Fedora Commons has already done a good job of building out the community and opening up communication. I would encourage you to continue with that work.
Besides that, one of the best things Fedora Commons could do now would be to focus on strengthening the brand. I still come across a lot of people who don’t understand how Fedora is different from DSpace, ePrints or Greenstone. The differences are pretty substantial. It would be good to publicize them up front. In particular, you should address, head on, the fact that Fedora is not an out-of-the-box system. Fedora is an architecture, and it is designed around the idea that storage, preservation, indexing, and security should be separated from user interface, metadata schemas, content models, and whatever management/re-use applications you might cook up. Fedora takes more work to use, but it gives you a huge amount of flexibility. Most importantly, it gives you far more opportunities to re-use and re-invent your content over the long term. By owning the reality [of what Fedora is and is not] you gain a lot of footing.
HatCheck: For Matt’s complete account of the “Unpacking Fedora 3.0″ workshop, and to follow his listening tour, take a look at the MediaShelf Blog.
