external parties (provided there was a way for third parties to point to ERA records unambiguously, e.g., via a unique identifier), or the ERA could hold and provide access to the overlays.1
The continued introduction of specialized enhanced capabilities by partners and vendors to support particular needs. For example, a third party or NARA might supply an interface that provides translations of popular audio recordings (e.g., the Nixon White House tapes) from the format in which they are archived into commonly used formats (e.g., Windows Media or Real Audio). Some of these capabilities might prove useful enough to be folded back into the ERA system proper.
Entrepreneurial innovation to provide value-added services, through either APIs or bulk download. Today, for example, third parties provide value-added access to information in the Securities and Exchange Commission’s Electronic Data Gathering, Analysis, and Retrieval (EDGAR) system. Similarly, commercial third parties currently provide enhanced access to old data of the U.S. Census Bureau.
Educational materials. Overlays onto NARA holdings could be a relatively low cost method of dramatically providing a new set of services that satisfy NARA’s educational mission. Educational publishers, as well as historians and social scientists, could develop online materials that draw on NARA holdings.
Such capabilities would allow NARA to tap at low cost the resources of third parties in order to provide enhanced services, help fulfill NARA’s mission, increase the perceived value of NARA’s holdings, and explore new technologies and approaches.
In allowing such capabilities to be built on top of the ERA, NARA will have to consider the problem of how a typical user distinguishes between results that are provided directly from NARA and those produced via a third-party extension. It would probably not be practical for NARA to certify third-party software or services. However, NARA may want to find ways of letting users know when they are using NARA versus third-party software and services, because the guarantees that NARA would make about the authenticity of what is viewed in the two cases would be very different.
Additionally, support for bulk download of records should be provided by NARA to allow third parties to retrieve sets of records against which to run their own full-content searches, perform data mining, and so forth. And an additional design issue that would facilitate third-party access is the adoption within the ERA of a persistent (i.e., stable and durable) record-identifier scheme.
Federation is a technique whereby a software layer is used to make a collection of relatively different systems—which are often controlled administratively by different organiza-
The overlay concept also has application internally within NARA. For example, Thomson West undertakes to rebuild the full index for the Westlaw databases only every 10 to 15 years. Between rebuilds, it augments the index with new information, some generated automatically, some manually. The system is structured so that a query collects information from all of these sources—the original index and the subsequent overlays. NARA could accept overlaid annotations from partners to add value to its holdings.