In addition to the guidance provided in chapter 2, the Implementation Checklist in appendix D may be useful in helping repositories plan their approach to implementing EAD.
Answers to some pointed questions about your repository's existing and potential clientele, as well as the ways in which your existing inventories are created and used, will help frame your decision making. Candor is required, but those who have previously implemented sophisticated descriptive standards or technologies, perhaps most notably MARC AMC, should feel at home with the process.
Your answers to questions such as these should prove helpful in determining whether the long-term benefits of EAD will be worth the effort involved in implementing it. In addition, give thoughtful consideration to the fact that for many archival repositories, adherence to standards is an increasingly important strategic goal when deploying new technology. Standards are seen as insurance that protects our investments in data creation and technology by making it possible to take advantage of a broader and more diversified marketplace, while also enhancing our ability to migrate data to future systems (an inevitable need). For decision makers such as library and archives directors who have adopted this strategy, the standards-based, community-endorsed aspects of EAD provide a convincing rationale for its adoption.
The investment in computer hardware and software that will be required is highly dependent on your local environment. Archives that already have the capability to produce finding aids in some sort of machine-readable form and to deliver information via the Internet are in a strong position to implement EAD encoding. On the other hand, a repository that has not yet automated its descriptive functions or integrated the World Wide Web into its public service operations would necessarily have to invest significantly in hardware and software in order to implement EAD in-house. Such institutions may wish to consider evaluating EAD's data structure for its potential to improve local finding aid practice while also investigating possibilities for identifying partners for a cooperative implementation venture.
The cost of authoring software ranges from trivial to somewhat significant, depending on the EAD encoding methodology selected. Similarly, software for publishing EAD finding aids also covers a wide spectrum of costs, depending on the approach taken. Chapter 4 and chapter 5 provide extensive information on scenarios currently available for both authoring and publishing.
While implementation of a new technology can perhaps most readily be viewed as a sinkhole for existing resources, it is also important to recognize the very real potential for conserving resources. By defining the elements of finding aids, and by suggesting effective means of arranging and displaying those elements, EAD can save you the expense and frustration of reinventing the wheel in the context of finding aid design, not to mention the cost savings in the public service context that result from improved finding aid effectiveness. These considerations are particularly potent for a new repository or for one seeking to professionalize existing operations, but even well-established archives are likely to find in EAD some strong ideas for improving the overall effectiveness of their descriptive practices.
In addition, numerous repositories have amply demonstrated that implementation of EAD has enormous potential for attracting new sources of funding. By providing repositories with a standardized approach to including finding aids (and linked digital surrogates of collection items) within digital archives and libraries, use of EAD has been shown to raise the public profileof repositories and to give potential funders confidence in the soundness of their investments.
Given that the introduction of any new technology has the potential to result in a wide variety of workflow changes, a broad review of your overall descriptive environment is highly worthwhile at the outset. For example, we may find that user access to in-depth electronic finding aids will allow us to rethink the length and level of detail of collection-level MARC catalog records.
Most centrally, however, EAD provides an opportunity to rethink the structure and presentation of your finding aids per se. Meissner has described how the Minnesota Historical Society evaluated its finding aids in the context of EAD's structure and elements before beginning markup.(35) This process included asking fundamental questions about the purpose of finding aids, their intelligibility to users when read without access to a reference archivist, their relation to collection-level catalog records in MARC, and their physical appearance. These aspects were evaluated in terms of the informational content of the finding aids, how users in the reading room perceived them, and how remote access might affect their use. Virtually every repository will find it necessary to undergo this process in order to determine how its finding aids map to EAD, both to identify relevant elements and to optimize their arrangement within a finding aid. Moreover, a thorough review of EAD is likely to suggest new techniques for inclusion and presentation of data that have not heretofore seemed necessary in a strictly local environment.
If your EAD implementation will be part of a multirepository effort or other consortial project, you must factor in the overhead attendant on decision making in a collaborative project. On the other hand, you are likely to have considerable outside expertise to draw upon as a result of working collaboratively, not to mention enormous cost savings due to sharing of hardware, software, and systems management costs.
Technology-based programs are always dynamic, subject to continuous evaluation and revision. Depending on the options chosen,(37) the level of technical expertise required to deliver finding aids electronically will range from modest to substantial. Archivists are resourceful people, and many early implementers of EAD have undertaken their projects with existing staff who have acquired new skills. Repositories that have access to in-house technical support staff, perhaps from another branch of the parent organization, may be able to tap into those resources, particularly within the context of bringing the archive into a broader digital library development effort. Purchase of support services is often a viable alternative, especially for routine activities such as setting up desktop computers and servers or installing and configuring common software such as Web browsers. SGML database search engines may, however, be less familiar to contract personnel, and so such services may be more difficult to locate or more expensive to acquire.
As usual, training options range from enrolling in formal classes or workshops to turning your more self-sufficient staff loose with written EAD documentation or software manuals. As mentioned earlier, the Encoded Archival Description Tag Library, Version 1.0 (38) is a key companion document to these Application Guidelines. Additional useful writings are available in print or on the World Wide Web, some of the most useful of which are listed in appendix G. The implementation case studies published in the American Archivist (39) describe a variety of collaborative self-education processes that are particularly effective in large or consortial environments. Whether or not you will engage in a collaborative systems implementation of EAD, joining forces with other local repositories to sponsor formal training workshops may help jump start your implementation process.
Graduate programs in archival studies have begun to incorporate EAD into their curricula on archival description and digital archives, and as a result, the job market will gradually be populated with applicants who come prepared to implement EAD. Such recent graduates will, nevertheless, often need training relevant to a repository's particular hardware and software environment. In some cases, graduate-level courses may also be available to working archivists through universities' increasingly popular distance education programs.
Most archives today create printed finding aids using word processing or database software. If one can continue to create finding aids in the same way and convert them afterwards, or substitute an SGML editor for the word processor, the impact on workflow may be minimal. Indeed, the net effect of such changes, once implemented, may actually be greater efficiency and less work overall. Meissner reports on the use of standard word processing templates, which at the Minnesota Historical Society reduced the time needed for keying container lists and the subsequent cleanup of data entry errors. (40)
As will be described in later chapters, numerous encoding choices must be made that will determine how labor-intensive this process is. One major consideration is the level of detail to be employed in tagging, an issue discussed extensively in the Library of Congress, Harvard, and Yale case studies in The American Archivist. Protocols developed range from minimal content designation focusing on finding aid structural elements, to experiments with richer markup that includes tagging each instance of proper names wherever they appear in a finding aid. For example, Lacy and Mitchell describe the numerous hypertext links embedded within Library of Congress finding aids. (41) As with many descriptive practices, greater effort at the input stage offers the potential for improved performance at the retrieval stage.
To determine which finding aids are most important or urgent for your repository to convert, consider which of them describe the following:
In addition to these issues, consider the fact that some finding aids offer more potential than others to enhance access, such as those whose content will yield greater rewards in response to a text search for key words or phrases. For example, searching a container list containing lengthy lists of names of correspondents or of photographic captions would significantly enhance name or subject access, while one chiefly comprising an enumeration of box numbers, volume titles, and span dates, such as are often used to describe certain types of organizational and governmental records, probably would not.
As for which finding aids will be fastest and easiest to convert, those initially created in some sort of digital format, whether using word processing software or a database management system, will almost certainly fall into this category. Two principal factors will influence how fast and simple that conversion is likely to be using a scripting tool or template:
When considering the feasibility of OCR, the legibility of printed data is a significant issue. A printed guide published using a clear typeface and without complex formatting, or a photocopy of one created on a high-quality, fairly modern typewriter, usually can be successfully converted to digital format using OCR. On the other hand, a finding aid created using an early typewriter or existing only in a poor-quality carbon copy or photocopy would almost certainly have be to be rekeyed. Rekeying may also be preferable for some finding aids that are otherwise suitable for OCR if their layout and formatting is highly variable or inconsistent.(42)
Hard-copy finding aids that are the easiest to convert exhibit the following characteristics: (1) clean ribbon or carbon copy with uniform typing and clear typeface; (2) uniform density of typing throughout the document with little or no handwritten corrections or annotations; (3) consistent formatting throughout the finding aid (uniform spacing, tabs, and or columns throughout the document, in particular throughout the listings of contents). Finding aids that are lengthy and fit into this category would be obvious targets for early conversion, since they will be quicker to convert than multiple shorter finding aids.
Hard-copy finding aids that are more difficult to convert exhibit several of the following characteristics: (1) inconsistent formatting (tabs or spacing, especially within listings of contents, are not uniform throughout, resulting in irregular columns); (2) typeface is hard to read because of the use of carbon copies or irregular typing and typeface; (3) multiple annotations or corrections have been made throughout the document, including handwritten attachments and listings appended to the finding aid.
The most problematic hard-copy finding aids to convert are preliminary or in-process listings, donor or dealer inventories, and card files. Generally such documents do not represent complete or otherwise fully satisfactory finding aids, and each repository will have to decide whether to expend resources to convert them "as is" or to wait until a complete finding aid can be prepared. Although converting such data may be anathema to most archivists, we are already beginning to find that many users (and administrators) have come to expect all our finding aids to be online, howsoever partial or imperfect, which may in turn cause a gradual shift in archivists' willingness to make imperfect finding aids available electronically.
Once your legacy data is in digital format, various characteristics of the data will affect how you convert it, as well as how difficult this process will be. As mentioned earlier, the consistency of the data will determine to what extent EAD tags can be inserted using macros or other automated processes, as well as whether a vendor would be able to successfully interpret the meaning of various types of information if you are considering contracting this work out. As described in more detail in chapter 4, some word processing software packages include scripting or macro languages that allow users to automate certain processes. In many cases, however, inconsistencies in formatting or descriptive practice will limit the extent to which repositories will be able to tag their finding aids in this manner. You should therefore probably assume that staff will have to spend some time doing manual editing of certain portions of your finding aids.
The level of tagging to be employed is another important consideration during the data conversion process, and reasons may exist to make different choices for converting legacy data than those you will use when encoding new finding aids. For example, you may be planning to reorganize aspects of your new finding aids to more clearly label specific EAD data elements (such as if your legacy data does not clearly separate biographical information from scope and content, or if details regarding acquisition and processing have been intermingled). It may or may not be feasible, however, to revisit all of your legacy finding aids at this level of detail during an EAD conversion project, and careful decisions will have to be made about which types of changes will and will not be made. A careful reading of chapter 3, in which the significance of particular groups of EAD tags is described in detail, will help you decide where to focus your efforts.
If the finding aids are to be electronically linked to an online catalog, linking data (in USMARC field 856) must be added to each corresponding MARC record as a pointer to the electronic finding aid. As described in section 126.96.36.199, the archive also may wish to supply an HTML version of each EAD finding aid for users with Web browsers that do not support SGML or XML. If so, a process for converting the EAD files into HTML must be developed and implemented.
Finally, many EAD implementers are offering explanatory materials on their Web sites describing what finding aids are, how they are used, and the methods provided for accessing them online. (43) This is an important service to provide to your electronic users to help ensure that they do not flounder for lack of a reference archivist to help them navigate your finding aids. The text of such online guides must of course be prepared, encoded, loaded, and maintained.
The involvement of regular staff is crucial to several aspects of this process. Planning and operational oversight are difficult, if not impossible, to contract out, although consultants may be very helpful at the start-up phase. Moreover, the contracting process itself generates administrative overhead. Certain skills are required: (1) a knowledge of the issues involved; (2) the ability to articulate institutional objectives and convert them into clearly measurable vendor deliverables; (3) familiarity with contract negotiations; and (4) an understanding of the dynamics of contract supervision, especially where quality control is an issue. A successful vendor-customer relationship requires both parties to have a clear and detailed agreement on their objectives and requirements. This would be particularly important in outsourcing EAD encoding, for example, given the wide range of choices that EAD provides for marking up finding aids. The options exercised in this area will directly and perhaps significantly affect the time required to encode a finding aid, and consequently, the cost of conversion.
The ways in which consortial members have been able to work together to plan for, educate themselves about, and implement EAD suggests a cooperative model that others may beneficially replicate. This approach will be particularly useful for planning, procurement of hardware and software, documentation, training, and technical support. As described in chapters 4 and 5, setting in place your approaches to authoring and publishing EAD finding aids can be both costly and extremely challenging. Sharing a joint implementation of these aspects of a project can therefore make the difference between feasibility and impossibility, particularly for smaller repositories. In addition, the added value that may accrue from the high public profile of a large collaborative project, particularly for a small archives affiliated with more visible institutions, should not be underestimated. And once a large institution or consortium has its EAD implementation up and running, it may be a straightforward process to take on more partners.(45)
On the other hand, it is important to recognize that consortial activity may carry other types of overhead. Broader discussion of issues and options in order to reach consensus invariably takes longer when more partners and interests are involved. Strong leadership and expertise is important to any such group; if the project leaders are disorganized, biased toward certain partners' interests, or otherwise ineffective, the project may be in trouble. Finally, if all partners to a cooperative endeavor do not take active part and make their required contributions, frustrations will inevitably result, since freeloaders ultimately cost everyone else time and money.
|Table of Contents|
|Home Page||Preface||Acknowledgments||How to Use
Aids in EAD
|SGML and XML
The Library of Congress