I.  Options for Addressing Finding Aid Discovery,
     Retrieval & Integration at Columbia

The current availability of MARC cataloging data for the Greene & Greene and Digital Aviador projects, along with functionality now available via CU Master Metadata File, suggest a practical solution to some of the issues & problems raised by the creation of archival finding aids at Columbia (see brief discussion of these issues in section II below).

By implementing what could be called "Index Records" either in CLIO or in the MMF and linking them to finding aids--either at the document level or directly to specific sections in a finding aid (identified with "targets" or "anchors")-- we could retain flexible and sustainable access to finding aid information as well as functional integration of finding aids within the larger electronic resource environment.

A summary description of this approach follows:

• finding aids would be "cataloged" in CLIO, using an index (or "access") level record rather than a standard descriptive catalog record; in some cases a finding aid would have only a single catalog record; in others there would be 'analytic' records for series or subseries within the finding aid;
  
  
• the "index record" would contain all necessary normalized name, subject and other headings, created according to standard cataloging practices;
  
  
• each "index record" would have a single link to the corresponding finding aid or to a location within the finding aid;
  
  
• "index records" would be loaded into the MMF, where image links would be created and maintained.  (NB: The master record for finding aid catalog records would in this case be CLIO, while the master listing of image or other digital object links would be the MMF.)
  
  
• finding aids would carry one or more internal "targets" or "anchors" which could be used for linking from CLIO or Web pages back to the finding aid;
  
  
• image (etc.) descriptive and structural metadata would be entered into the MMF for all images related to a finding aid (or similar digitization project), and the image stored according to guidelines (to be) established for CU's digital image repository;
  
  
• image links would also be incorporated as appropriate directly into the finding aids, but these would be pointers to the descriptive & structural metadata stored in the MMF;
  
  
• the MMF would be used to generate project and other types of Web listings and displays based on the access points in "index records"; these Web listings and displays could point users either to the finding aid or, if needed, directly to the images themselves;
  
  
• for projects other than G&G and Digital Aviador--where standard cataloging was already completed--finding aids would be given new index-level cataloging in CLIO (or possibly directly in the MMF depending on the nature of the project).
  
  

II.  Issues in integrating archival finding aids into
    Columbia's metadata environment

A.  Discovery, Retrieval, Integration

Effective 'discovery' and retrieval of information in finding aids may be difficult at Columbia because we currently lack:

• a technical infrastructure for indexing or searching individual finding aids or images linked to finding aids other than standard keyword indexing
  
  
• a strategy for cross-document searching of multiple finding aids
  
  
• a mechanism to integrate indexing or searching of findings aids with related materials not usually described in finding aids, e.g., material cataloged in the online catalog, local & remote online text archives, other digital library project data
  

B.  Data Maintenance & Technology Migration

Maintenance of finding aid information is currently difficult because:

• normalized access terms embedded in archival finding aids (e.g., names and subjects) cannot readily be updated or managed with existing bibliographic or other tools
  
  
• there is currently no single repository of archival finding aids at Columbia which would allow for effective management of the components (such as DTDs, CSS stylesheets, XLS stylesheets, various entity and rules files, etc.) necessary to support the current EAD/SGML apparatus or the necessary migration to XML, XLink, XML schemas, etc.
  
  
• there is currently no implementation of a standard "digital object lookup" or URN scheme at Columbia that would allow incorporating persistent links to images or other external resources into finding aid documents