[MOA2-WP:18] NYPL comments
pellis (pellis@nypl.org)
Thu, 30 Apr 1998 13:44:39 -0700 (PDT)
The NYPL group discussed the white paper Monday, April 27. We focused
primarily on the two new premises of this project--the object-oriented
approach and the structural metadata.
ARCHIVAL COLLECTIONS
The white paper takes a rather unarchival approach to archival
materials. All materials are treated as objects belonging to a class
rather than as elements of an archival collection. So, the white paper
does not address the collection as a whole or the relationship between
objects in a collection. Belonging to a class of objects should not be
emphasized at the expense of losing valuable information about
interrelationships between objects in a collection.
Provided that we strike a balance between classes of objects and
collections, we will need to see more carefully defined classes of
objects, e.g., "ledgers" seems to indicate any type of bound financial
material. This may obscure the function of a ledger, its relationship
to other types of financial records, and the way in which a researcher
would use these records together.
DESCRIPTIVE METADATA
Since the white paper doesn't address collections, it overlooks the fact
that finding aids are metadata about collections, not about individual
objects or classes of objects. They do not provide the "fine-grained
access" or the depth of detail needed for the analysis of objects and
sub-objects.
On the basis on other projects with which we have been involved, we
didn't find the focus on item-level description unusual. We do find that
item-level description, particularly to the depth proposed for this
project, is not scalable to mass conversion of archival collections.
STRUCTURAL METADATA
Since the focus of the white paper was on the internal organization of
objects, we looked more closely at the proposed fields for structural
description. The structural metadata elements proposed largely represent
existing document structures (e.g. books) or are rather vaguely defined.
Since the proposal describes goals for user interaction with the objects
more broadly, these fields will need to be discussed at greater length.
We are asking the input of our internal metadata task force to take a
look at these elements, too.
DELIVERY
The object-oriented approach in combination with the structural
description presents a very promising way to deliver these materials,
but the details of how this is to be achieved are sketchy. The proposal
seems to waver between creating programs to act upon the data (like Java
applets) or simply creating the descriptive and structural metadata that
will then be acted upon by the individual institutions' server-side
apps.
We were relieved to see that data may be stored in USMARC, relational
tables, or SGML. Since we use both ODBC compliant apps (Cold Fusion and
Informix) and SGML-to-HTML apps (DynaWeb) on the server-side to deliver
our digital library collections. We have also done some mapping between
metadata formats to leverage existing (often collection-level) MARC
records for our digital library collections--something the white paper
did not adequately address.
NYPL SELECTIONS
We are in the process of revising our selections in light of the reduced
funds. We have tentatively decided to eliminate the NY Central, Long
Island, and Pennsylvania Railroad material from our proposed selections.
This leaves the Jackson diaries and correspondence, accompanying
stereographs and prints, and the Stanton journals, manuscripts, and
photographs.
Some of this material will need to be transcribed, so any information
about vendors or other approachs to transcription would be appreciated.
Here's what we have so far (we are still reviewing the collections) in
terms of the classes represented in the project.
*Continuous tone photographs
Jackson stereos (to accompany diaries), Stanton photographs
*Photo albums
none
*Diaries, journals, letterpress books
Jackson diaries, Stanton journals (field notes, reminiscences, and
manuscript--"The River and the Canyon")
*Ledgers
none
Correspondence
Jackson