Mellon funded Museum Data Exchange project

You may have already seen a press release, but in case you haven’t, I’ll first give you the gist, and then some insights into the work already underway.

With the generous support of a $145,000 grant from the Andrew W. Mellon Foundation, RLG Programs will gather a select group of museum partners to accomplish the following:

Creating a low-barrier / no-cost batch export capability for CDWA Lite XML out of the collections management system used by the participating museums (GallerySystems TMS)
Modeling data exchange processes using the Open Archive Information Protocol for Metadata Harvesting (OAI-PMH) at the participating museums
Creating an aggregation of museum content within OCLC Research for analysis
Discussing the evidence about the relative utility of the aggregation with stakeholders from the museum, vendor and aggregator community

It’s a tall order for a project with a 15 months lifespan, and that’s why we’ve already gotten started! On January 28th and 29th, representatives from the five institutions participating in phase 1 (Metropolitan Museum of Art; Museum of Fine Arts, Boston; National Gallery of Art; Princeton University Art Museum; Yale University Art Gallery) gathered at the Met to discuss the functional requirements for a data extraction and publication capability. Most of the participants had already dipped their toe into the water – in one way or another, they were trying to locally build a mechanism to generate CDWA Lite XML records. From this vantage point, the grant becomes an opportunity to jointly develop a tool which can be used by the entire community, rather than making redundant local investments which only benefit single institutions.

Of course the biggest challenge in this endeavor is the variability with which museums implement collections management systems. Each museum uses the data fields offered by TMS in a slightly different way, and oftentimes, there is even variation within a single installation as different departments use different guidelines for data entry. The Met serves as an illustrative microcosm for this challenge: collections are described in 20 separate installations of TMS, and each instance thinks about museum data in a slightly different way. At the Princeton University Art Museum, different mappings will be required for data created by different departments. As a consequence, a tool which aspires to be usable by any TMS installation has to provide a mechanism not only to tailor the mapping from TMS to the standard output (CDWA Lite XML in this instance), but also to allow users to apply different mappings to different sets of data within that installation.

During our meeting in New York, we agreed that one way to reduce the complexity of creating these mappings would be to make them shareable. As a baseline, we hope that our tool will ship with a default mapping which project participants determine takes them the furthest towards a meaningful output. Beyond the default mapping, museums will be encouraged to share their customization of data mappings on a field-by-field basis. Maybe your installation of TMS thinks of creators just like the one at the Yale University Art Gallery, but you use dates just like the MFA Boston. Combining these field-by-field mappings into a profile will allow museums who are new to the tool to generate a satisfactory CDWA Lite XML output with greater ease.

As I’ve been quoted as saying in the press release, while we are focusing on TMS for the initial testing of the tool, nothing in the technologies we are thinking of deploying would prevent the extract from working against other databases. As a matter of fact, both the Met and the MFA Boston were speculating about initially using the tool against a non-TMS consolidated datastore of records.

I’ll keep you posted on this effort as we continue our work on this grant. We’re off to a good start for phase 1!