- Current Columbia Approach to Finding Aid Creation and Publication
At present, Columbia has no consistent approach to creating and
publishing finding aids across the units of the Libraries or university-wide.
The Rare Books and Manuscripts Division has developed style guidelines
for presentation of their finding aids, although this has not
yet been extended to other divisional units creating finding aids.
Individual archivists and curators within different organizational
units have each developed their own personal approaches to creating
electronic finding aids using chiefly ProCite and FilemakerPro,
which are also used to generate HTML output for the web. We
are experimenting with the use of XMetal & XSLT in one division
(Avery). Columbia has no single master archive
of finding aids; we provide no indexing across finding aids; finding
aids are not integrated into other non-archival web resources.
Our goal over the next year to provide a more coherent technology
approach to finding aid creation and management.
- Finding Aids In Context
From Columbia's vantage point, the task of creating and providing
access to archival finding aids presents a number of issues and
challenges, not the least of them conceptual. In
developing a coherent strategy in this area we feel it important
to recognize that finding aids are not ends in themselves. They
are, in the broader context of the digital library, a way of providing
a certain, limited type of intellectual access to a single archival
collection. The finding aid model was developed in the paper
and print world and remains, even in its EAD / XML formulation,
solidly based there. As a metadata structure, it diverges significantly
from the data element / database model developed for other types
of resources; as an end-user presentation model, it resembles
nothing so much as the notebook with index tabs that it has replaced.
More significantly, the tendency to focus on the finding aid alone
ignores equally important challenges facing archivists as well
as those attempting to develop workable digital library strategies,
namely:
- Archival collection processing & management
- Flexible & effective end-user presentation
- Intellectual integration with other non-archival digital
library resources
- Functional integration with key digital library infrastructure
components
The lack of available tools and solutions in these areas makes
the already enormous workload of processing and providing access
to new collections even more overwhelming. In that context
it seems to some archivists that the task of just creating and
publishing electronic finding aids for all of their collections
is already more than they can aspire to.
The specific task areas listed above are described more fully
in the following sections. It is clear, however, that these areas
will not all be addressed collaboratively or probably even within
a single institution in the near future. Still it is important
to begin developing requirements and creating persuasive data
models for these areas so that we can plan effectively at the
institutional level and, interinstitutionally, develop consistent
and compatible solutions. Some types of solutions (e.g.,
integration with local infrastructures) will always need to be
locally developed and managed; with others we may be able to work
collaboratively and perhaps even interest automation vendors in
providing the toolkits and software support that could substantially
reduce archival control costs for institutions.
- Archival Collection Processing & Management
Given the enormous size and number of new and unprocessed collections
that libraries and archives have acquired and will continue acquire,
tools are urgently needed to support core tasks such as:
- Intake & processing new collections
- Adding material to existing collections
- Conserving materials
- Digitizing materials
- Managing permissions & intellectual property rights
-- including those of digitized materials -- often at the
item level
- Tracking various kinds of collection use, including: on-site
consultation, publications based on the collection, loan of
collections for exhibitions
The work carried out in development of the EAD
standard has already contributed a large chunk
of the analysis that would be needed to create
a collections management data model; non-archival
digital library efforts have supplied some of
the other important components. What is
needed is for the whole model to be fleshed out
so that the individual components, including finding
aids, can be shaped to interoperate with other
functional pieces of the electronic environment.
- Flexible and Effective End-User Access
The traditional finding aid presentation is that of a continuous
text that provides information at the collection, series, subseries,
container, and sometimes item level, usually presented in an order
dictated by the physical organization of the collection.
In an electronic environment this model may not always be the
best way to present such information to users. In some online
archival projects the "continuous text" approach to
finding aid presentation has in fact been replaced by a severely
menu-driven, hierarchical display; in others, presentation is
chiefly in the form of database-style query responses. Neither
of these alternatives -- on their own -- is necessarily an improvement
over the traditional 'tabbed notebook' display.
What is needed is the ability to easily produce different kinds
of presentations for different types of collections, and sometimes
for the same collection. Checklists, browsable lists of
key facets of the collection (e.g., projects, publications, names,
chronologies, geographic content) or special online displays of
significant collection subsets (e.g., architectural drawings displayed
in conjunction with photographs of the completed buildings), will
often be useful as supplements to or even replacements for the
"notebook" format. This is particularly true of
archival projects having a digitization component, where locating
digital content should not necessarily require users to 'drill
down' into the finding aid, and where some of the archival content
will need to be repurposed for other presentations, e.g., an online
exhibition based on the same collection.
- Intellectual Integration with Other Digital Library Resources
The EAD finding aid format as usually implemented does not lend
itself to functional integration with information about other
archival collections or other types of library and museum materials.
As an obvious example, a collection of architectural
drawings may have excellent scholarly studies available in the
libraries' online catalog, or closely related materials elsewhere
in the local or national digital library -- but these facts may
never be known to someone viewing the finding aid.
For these reasons, the metadata aspects of finding aids need to
be able to be easily integrated, merged, indexed and presented
along with other types of metadata. Moreover, metadata associated
with the finding aid needs to be able to be managed in the same
way and with the same tools as other non-archival metadata.
- Functional Integration Digital Library Infrastructure Components
Since many archival collections are candidates for digitization,
at least selectively, the display and access management of digital
content must also be accommodated in one way or another. The finding
aid format (and the EAD / XML standard) does not lend itself to
these functions.
Structural metadata, relationships to digital object repositories,
relationships to access and intellectual property controls, association
with administrative metadata, cross-searching of normalized headings,
etc., must all be available to and interoperable with finding-aid
type information in the digital library environment.
The fact that most archival collections may never have digital
content or require the type of detailed (and expensive) integration,
management and end-user access functionality described above doesn't
change the fact that some will and already do. And
collection-based digital library projects in the future will almost
certainly need the infrastructure to support this greater degree
of functionality. The expanded archival data model
thus needs to flexibly support both simpler approaches as well
more complex & fully functional implementations.
- Conclusions & Recommendations
- The EAD movement has proven itself as a powerful force for
improving the documentation of archival collections. However,
the early decision to take an SGML "text-encoding"
approach -- as opposed to a metadata / data element / database
approach -- has shown itself to be limiting, particularly
in the toolsets available for data collection & management
and web publishing. The SGML choice has in some respects marginalized
archival control, separating it from the mainstream of digital
library development and support. It may be that XML
will ultimately provide the tools needed to create and publish
finding aids in the way the EAD developers originally envisioned;
but the technologies for it have not yet matured, and it is
still unclear how and whether other digital library infrastructures
will migrate to XML. The most nearly suitable tools for finding
aid creation, publishing, indexing, etc. in an EAD context
(DTD-aware text editors, XMetal, XSLT, DLXS/XPAT, Tamino)
are not part of most institutions' existing digital library
infrastructure and are, variously, not scalable, not affordable
or not yet mature enough technologically for most institutions
to justify new investments in. For the foreseeable future
it is urgent that we reconnect and integrate finding aid creation
& management with the robust, existing tools that have
proven themselves over time -- most specifically with database
technologies.
The largest finding aid publishers and aggregators (e.g..,
California Digital Library, RLG) have already made heavy use
of databases to support both data collection as well as web
publishing of information derived from finding aids.
An obvious and useful direction to pursue would be for
the archival / digital library community to develop a standard
SQL schema that parallels the EAD DTD. This would allow
data to be moved from EAD / XML finding aids into databases
and back, while providing the greater stability & functionality
of databases for managing and publishing digital library information.
If this archival SQL schema were developed & documented
via a case tool such as Popkins System Architect or Rational
Rose, it could be exported as needed to different SQL database
platforms.
This single step would immediately allow more effective integration
and use of finding aid information in the short term, and
would provide a more flexible and fully-supported testbed
for the development of new archival tools and strategies in
the future.
- The RLG archives database could play a much greater role
in integrating archival resources nationally
if it were more proactively managed, developed
and publicized. Cross-collection
finding aid searching is difficult at
the institutional level and virtually
impossible at the national level without
a database such as as this. We
should ask RLG what its level of commitment
and support for this system, and work
with them on improvements and expansion.
A fully-supported national archival
database might make it unnecessary for
smaller and mid-size institutions to make
heavy local investments cross-collection
search and retrieval.
|
|