Bruce D'Arcus wrote:
>As for my own standpoint I've sort of come to the conclusion that our
>citation needs are rather different than the needs of libraries. I am
>trying to push that OpenDocumen include DC and Qualified DC out of the
>box, and provide a framework for predictable (RDF) extension.
There is one extension that I would love to see in DC or in any other
metadata format that will be encoding journal-type items: enumeration
and chronology. The MARC format does this well (logically), but it is
hard for people to grasp because of the MARC structure, I think.
Essentially, one of the great complications of serial publications are
these hierarchies of volume, number, etc. What's complicated is that
there is a near-infinity of patterns that has been used in actual
practice. Some journals have parts that are hierarchically superior to
volume, others have parts that are subdivisions of volume, some journals
go straight to number, and some have no numbering at all. A fairly
simple structure could encode many (probably not all -- serials can be
very weird) of these by allowing any number of levels as long as they
are 1) in order and/or 2) named. So I'm thinking that a structure like:
Which essentially translates to: Volume 5, Part B, Number 12, December,
1996, but if your journal had the part before the volume, then part
would be level 1, etc.
This should also be able to accommodate journals with un-captioned
levels, any number of levels, lack of chronology, journals with only
chronology, etc. This is what the 85x and 86x fields in the MARC
holdings record can do. It will probably need another data element or
two for exceptions (i.e. for index volumes), but we should be able to
come up with something that works pretty well without much effort. Or am
I totally deluded?!
Karen Coyle / Digital Library Consultant
[log in to unmask] http://www.kcoyle.net