… and I say that not just because Jon and I have been working with the Project on their metadata planning. It’s an excellent paper, and we’re delighted it’s out, so we can tell people about it, and encourage them to read it and talk about it. We’ve been talking up the XC Project as one that has been doing important work and moving forward in ways that reward our close attention. But until now we’ve found it hard to find something to which to refer interested parties, given that so much about what XC has been doing has been happening under the surface of the pond. Now we won’t have that problem. Go look at the paper here.

Jennifer generously mentions the inspiration of our ideas—as reflected in one of our grant proposals (which wasn’t funded and isn’t publicly available). We’d like to point out that many of the ideas in the proposal are available in previously published papers, all of which are linked from this site.

Takeaways from XC

Jennifer rightly and effectively presents the practical decision points in the XC project and documents them well. But there are bigger issues underlying those decisions, and it’s important to note how they fit into the larger task of building services in an environment as fluid as the one in which we find ourselves.

“It is also important to understand that the XC Schema represents a purely practical rather than a theoretical approach to using metadata. While much interesting work is taking place in the library community to develop theoretically consistent metadata frameworks, models and carriers, the process used to develop the XC Schema should not be confused with these start‐over‐from‐scratch‐and‐do‐it‐right approaches. In some ways, the XC Schema represents exactly the opposite. For XC, we have attempted to manipulate and make use of the metadata in various schemas that is available to us, and to use this metadata in a way that it can be shared with others. While ours is not a theoretical approach in itself, we believe that the work that we are doing can inform the work of those who are developing more theoretical approaches to the future of library metadata. Our experiences should provide a reality check regarding the feasibility of retrofitting legacy metadata into a new theoretical framework. Much of the effort needed to retrofit library metadata for use in new applications alongside metadata that originated in other schemas is likely to require considerable experimentation. The XC Schema is designed to embrace and facilitate that experimentation.”

Despite Jennifer’s emphasis of the practical solutions (and the fact that she has a notion of experimentation that is foreign to most librarians), there are some important theoretical underpinnings that she illustrates in the paper, and I wanted to hit on a couple that I think are important.

The benefits of “dis-integration”

It’s ironic that libraries have been urging their vendors for years to integrate their platforms better, and now we’re taking a quite opposite approach. Jennifer nicely describes how this works in XC, and the rationale behind those decisions. Many in libraries have come to the conclusion that discovery services needn’t be bolted to data management services, but XC goes several steps further, and unbolts data management from the ILS as well.

“Dis-integration” is critical in bringing libraries out of the current environment of tightly bound library systems built on MARC and depending only on one another for data. Libraries today have too many silos of data for that strategy to endure: IRs, projects of all kinds, commercial partnerships, etc.

Iteration vs. Perfection

I was talking to a colleague new to libraries recently and describing how data was moved from one platform to another in the days when I was still working in academic libraries. Back then, we spent almost a year planning the “migration” from NOTIS to Voyager, and had contractually specified two test loads of our data into the Voyager test system we had up. These loads were then tested extensively, using specifically chosen test conditions and records (we had our own test plan, which was extremely detailed and fine tuned—we didn’t trust the vendor at all). After each load we would specify changes that had to be made before we could do another test load. The third load was the real one, it was final, and all the data had to be as perfect as we could make it, as we had no other chances. For months before the loads we had students sitting in front of computers making changes we couldn’t make programmatically and programmers slogging through the ones we could, knowing that it would take years to re-create the knowledge and tools to do them in a new system. They did the last changes about 20 minutes before the data was extracted for the “real” load. Stress, anyone?

Consider the difference in XC. The normalization and transformation services are independent, they maintain their own data stores, and they’re designed to operate independently. Data services can be tweaked as experience and changed goals dictate, and data can be moved through the services any number of times, until the result is optimal, without interruption of any user delivery services at all. It’s an ideal strategy for an environment where nothing is stable, change is perpetual, and nobody can afford to wait until things settle down (they may never, you know).

The XC strategy also supports the construction of an environment where many entities can build usable services and the expansion of overall service capability (and the cost of creation and maintenance) can be shared broadly.

Taking the Leap

One of the most important things that Jennifer’s paper illustrates is how libraries can make the leap from legacy data and systems into the future of library data. A lot of people are scared by this — they figure that there’s no way to make the leap unless we know more, standards are more complete, agreements are made. XC has provided real leadership here, by showing the value of jumping in and trying things out. They are showing far more leadership than the institutions (you know who they are) we have always depended upon to lead us forward.

The only possible response to this effort is close attention and engagement. Take a look at the supporting documentation Jennifer provides, engage her in conversation about this work, and take inspiration from her willingness to try out new ways of doing things.

We’re all very proud of you Jennifer—keep up the good work!

By Diane Hillmann, January 13, 2009, 4:52 pm (UTC-5)

Add your own comment or set a trackback

Currently 1 comment

  1. Pingback by FRBR vs. FRBR-ization | Metadata Matters

    […] eXtensible Catalog folks have discovered as they build the services to transform MARC into RDA (see my post on Jennifer’s paper for more about that) transforming MARC to RDA represents a fundamentally different set of problems […]

Add your own comment

Follow comments according to this article through a RSS 2.0 feed