I am assuming at the start that we will agree that some type of encoding standard and a prototype implementation of that standard are the two major products that will advance our objective. That is, we need to produce a structure and a working demonstration system and not simply theoretical or descriptive papers on the concept.

In order to develop such a product, we need to address a number of issues dealing with the existing environment for archival authority information and its relation to other archival descriptive products; the state of acceptance, knowledge, and experience in the archival profession of the concept; and the likely path to adoption of the approach as a standard and as a practice. My comments here draw heavily on both the MARC-AMC and EAD experiences and the roads they traveled on their way to acceptance by the community. I am not convinced that there are substantive differences in this project from any other standards development project, other than the degree of consensus in the community about the inherent value of the approach. So, in many ways, this is simply an effort to articulate that process for this particular issue and try to define specific implications for our work.

The comments also are based primarily on a US perspective and the experiences and state of the archival community in other countries may differ significantly and suggest modifications.
 

Define and confirm the concept and show it working

While the idea of archival authority information, as defined in the first discussion paper, has been discussed by archivists for over 10 years, recognized in an international standard (ISAAR (CPF)), and implemented in a small number of existing systems, it is not clear that it is recognized as an integral part of archival description by more than a small number of institutions and archivists. Most archivists will readily subscribe to a descriptive practice that includes the recording of biographical or agency history information in traditional ways - as narrative text in a catalog record or finding aid - but relatively few would understand the benefits of recording it as a separate but linked entity or as information that can have its own structure and content designation. One of the major tasks we face in having this concept of archival authority information adopted as a standard and a practice is to convince a critical mass of archivists that it is beneficial, feasible, and economic to treat authority information in this way and to demonstrate that the tools are available to make it happen.

Most of the discussion in the profession thus far has focused on administrative histories of organizational units, where a variety of rationales specific to the mission and responsibilities of the archival repository (such as citizens' access to information about their government) and to the method of organizing records (multiple series, each with its own descriptive record, originating from the same organization) make the value of structured authority information self-evident. For materials, such as personal papers, however, the value of "authority copy" that can be used in multiple places is not as obvious. Nor is there the same mission on the part of the repository to record and disseminate information on a person's life and activities as there is for an archives to do so for a governmental agency. {Bright Sparcs is an exception since it focuses on individuals and their characteristics.) We need to consider carefully the arguments and benefits we want to advance for extending the approach beyond organizational entities.

We should also consider the application of the concept to other non-records creating entities, such as functions, activities, and artifacts. The importance of architectural research on the Yale campus, for example, strongly suggests that an authority record standard that encompasses structured and content designated descriptions of buildings would be extremely valuable and, in fact, such information already exists in book form and is heavily used. Similar cases can be made for functional, geographic, and other entities or concepts that are documented or influence the creation and use of records. One of the goals then is to determine whether it is practical and useful to establish one authority encoding standard for all types of entities or to develop separate ones for each specific type.

The existence of working prototype systems that demonstrated how effectively catalog-level information and finding aids could be searched, retrieved, and exchanged was key to their acceptance by the archival community. In both cases, there was considerable skepticism about the value of the undertakings and the amount of effort needed to reorganize thought processes and workflows to accommodate these new standards (and some still remains). Having a working model with live information that can be reviewed and tested will be essential to acceptance and involvement by the profession.
 

Accommodate existing universe

Since adoption of this practice will involve a change of methodology, at least for many in the US, the benefits must outweigh the costs of reorientation and restructuring of information systems. As with MARC-AMC and EAD, the proposed standard cannot be overly prescriptive in terms of required content designation and structure. It needs to accommodate the existing universe of authority information so that it can easily be migrated into a more robust system that treats it as independent, linked data without requiring extensive analysis and recoding.

In addition to the technical accommodation of existing practice, the proposed standard must also be careful to identify and address the full range of content and structure that are already being used. While there are few instances of highly structured archival authority information in place, the content that could be specifically designated within existing information products is extensive. Since much of this information is embedded in narrative texts, it will require analysis of these texts to identify content that is commonly recorded and that may warrant either structure or specific content designation. This parallels the work that Elaine Engst did prior to the development of MARC-AMC and that the Berkeley team did to "distill the essence" of finding aids for the document type definition.
 

Encourage better practice

At the same time that it accommodates existing information products, however, the standard also needs to be sufficiently rigorous to encourage better practice in analyzing, structuring, and content designating authority information. It can best do this by a combination of a detailed standard that reflects a complete set of relevant data elements and clear examples of how they are used. To the extent that such structured and encoded examples can be compared to the looser narrative structures with which most archivists work, and the benefits of such structuring for retrieval and comprehension purposes shown, the standard may gain more adherents.
 

Prototype that demonstrates linkage and navigation

Also critical to the success of this effort will be a working prototype of authority information linked to both standard catalog records and encoded finding aids. At a minimum, this prototype needs to show the feasibility of the concept such that one can easily and intuitively navigate amongst all three types of archival descriptive information. The prototype needs to provide access to catalog records and finding aids in existing systems from authority information rather than requiring a migration of the former into a custom-built system. In this way, it is similar to the capability of linking MARC catalog records that already exist in national utilities or local systems to EAD-encoded finding aids.

Again, to the extent that such a system can be compared to existing architectures and methodologies and shown to be more intuitive and effective, there will be a better chance of convincing archivists to adopt this approach. If user studies are built into the project, then the prototype should be built in ways that the data needed for such studies is readily available. The design of user studies should take place concurrently with the specification of the prototype system with a good sense of what will be investigated.
 

Review and standards process

The encoding standard and prototype should be constructed in such a way that they will be readily available for review and comment by the international archival community at appropriate stages. The process by which the FINDAID document type definition became EAD is a model for the type of review and consultation that is needed. A call for the contribution of existing archival authority information, coupled with direct invitations to repositories who are known to have caches of such information or represent certain archival sub-universes that may have different needs, can publicize and educate the community to the approach (as well as assembling the needed information).

Similarly, a structured process of review and comment with a plan for turning the process over to an established professional standards group for final development, promulgation, and maintenance may be a goal we wish to articulate and aim for from the start. Presentations at professional meetings, articles, and similar consciousness-raising and education activities should also be explicitly planned for.
 

Summary of the characteristics of the project product:

Convincing demonstration of the utility of the approach.

Demonstration of tools to make implementation practical and cost-efficient.

Ability to link technically to related descriptive products in existing systems.

Ability to accommodate existing practice (within reason).

Encouragement of the development of better practice.