On Mar 23, 2005, at 1:58 PM, Riley, Jenn wrote:
> I'm reading an assumption here (and please correct me if this is an
> unfair characterization!) that the FRBR structure is completely
> hierarchical, and that each level from
> Work->Expression->Manifestation->Item can be nested. The problem is
> Manifestation has a many to many relationship with Expression - a
> Manifestation may contain multiple expressions of works, as in tracks
> a CD. So there needs to be some mechanism for connecting a
> to multiple Expressions. Only having pointers "downward" (excuse the
> term implying hierarchy) would then require redundant Manifestation
> What would be a good METS way to represent these sort of relationships?
> I'm not thinking of any off the top of my head.
Two nit-picky points, and then an answer to your question:
1. A manifestation actually cannot contain an expression. An
expression is a realization
of a work; a manifestation embodies that particular expression. So, by
reading, the expression-to-manifestation link is one-to-many. An
can have several different manifestations, but any particular
the embodiment of only one expression.
2. Thus, to my mind, a CD containing multiple tracks isn't a case of a
(the CD) containing expressions; it's a set of manifestations gathered
in single physical package, which might or not count as a
on your point of view.
That being said, there's nothing very difficult about having a single
that is included in more than one higher-level manifestation. Again,
is the <mptr> facility. Create a METS object for the individual track.
A METS object
for the CD can then use <mptr> to link to the track; a different METS
representation for larger piece that that track comes from?) could also
to the same track METS file via an mptr.
Digital Library Development Team Leader
Elmer Bobst Library, New York University
70 Washington Square South
New York, NY 10012
[log in to unmask]