RFC 1691 (rfc1691) - Page 3 of 10

The Document Architecture for the Cornell Digital Library

RFC 1691               CDL Document Architecture             August 1994


   6. Low-resolution "thumbnail" images of each page must be stored to
      facilitate browsing and sharing of data.

   7. The architecture must support distribution of files so that
      similar files may be stored together, permitting optimization of
      storage use and performance.

   8. The architecture must support documents that are composed of
      references to all or part of other documents.

   9. The architecture must support document components which are
      stored on separate servers distributed across the network.

   10. The architecture must support not only an hierarchical structure
       for each document, but the ability to define multiple views of
       each document.

   11. The architecture should accept, rather than dictate, directory
       structures in which documents will be stored.  This will permit
       documents created in other ways to be added to the Digital
       Library simply by adding database information rather than by
       copying or moving files.

Document Architecture Description

   A digital library consists of a Digital Library Server, networked
   storage, and a referencing database.  A single digital library will
   contain one or more collections.  Each collection will contain one or
   more documents.

   The referencing database allows searching for documents by author,
   title, and document ID.  In the current implementation, the
   referencing database is a relational SQL database, and each
   collection is  epresented by a table in the database.  It is planned
   to migrate to Z39.50 database searching as the preferred method, as
   this protocol has been established as the standard for library
   applications.

   Authorization will be primarily collection-based, although the design
   will permit authorization checking at any level down to the
   individual file.  Notification would come only when the patron
   attempted to open the document or access the particular component.

   Each document consists of three components: the logical structure;
   the physical references; and the data files.






Turner