05/04/2000 03:03 PM

Glossary for Reference Linking

Note: in the following substitute the word "creation" for the word "work".
The IFLA definition of "creation" includes works, expressions, manifestations,
and items.  References and citations are examples of works (they have authors, e.g.)  
while postscript files in an archive are items (they are instances of a 
creation, formatted in a certain way.)

-  Reference
  1.  An ambiguous natural language string in the body of a work to
      another work.
      Example: "that paper Donna wrote in 1975 about reference linking".
  2.  An ambiguous but structured reference to a work.  Example:
      (Lagoze, 1998) when he wrote 5 papers in 1998.
  3.  An unambiguous structured reference, but not linkable.
      Any reference to a 1980 CACM article.
  4.  An unambiguous, structured reference that is linkable.
      Any reference to a 1999 CACM article.

  A reference is to a work and is contained in a work.  A work has a
  fixed number of references for all time.

-  Citation
    Something that a work may acquire over time.  A work acquires a new
    citation when it appears as a reference in a new work.  Usually this is
    discovered by analyzing an item of the latter.

-  Work - similar to what work is in IFLA.  It is paper written by
    somebody, for example.  It may or may not appear online.  It may
    in fact no longer exist physically or digitally.  But it still has
    a title, authors, year published, etc.

-  DID, the unique identifier for a work (borrowed from ResearchIndex)
    is the URI for papers.  The identifier for a work may be found by
    specifying some bibliographic data for the work.

-  Location: the URL where a copy of a work may be found online.

-  Link: a reference or a citation.  A cite-ref.  

-  Bibliographic Data - stuff like title of this paper,
   authors of this paper, year published.  Sometimes also called "metadata".
   
- OAMS: Open Archive Metadata Set.  A minimal set of metadata expressed in XML.

- Surrogate: In this particular API, Surrogate means a digital object which contains
metadata and bibliographic data for an archive item.  Think of it as a java class 
with fields and methods.  In fact, if we go the FEDORA route, all the surrogates are
persistent objects, stored in a digital object repository, and accessible via CORBA.
If we don't use FEDORA, then think of the Surrogate existing in a series of databases,
holding bibliographic data for each item in the archive.