Printable Version

Requirements for Metadata Registry

The metadata registry project is a complex project.  Discussion to this point has focused mainly on what the content of the catalog records be and, to a lesser extent, on how those records will be put into the database and who will produce them and to an even lesser extent on how these records will be searched and exposed to applications other than search engines. 

It has also been the case that until March ’03 all of the work on this project was done over the course of last summer and resulted in the creation of a prototype (http://www.lib.utexas.edu/dams) that demonstrated certain concepts that are fundamental to this project:  1-that catalog records will be produced by staff in departments other than the library; 2-that metadata will be stored apart from the digital objects it describes; 3-that external applications will be created that will use the metadata in the catalog; 4-that this project will produce news ways of working among those involved in the production and management of digital content for instruction, research, exhibition and public use at UT.  As originally conceived it will serve as a centrally organized and managed catalog containing information about a wide variety of information objects produced on campus.  The objects themselves will be distributed across campus.  The registry will become a tool that will provide an authoritative source of information about digital objects on campus that will enable developers, end-users, managers, content producers the ability to find authoritative information about the existence of digital objects more quickly and more easily  This approach could be called a “union catalog” and has proven very successful in the library world. 

While the database will be centrally managed the process of creating the catalog records can be distributed to the content providers through a cataloging module that will be built as part of this effort.

The registry as management tool will simply store information about digital objects.  Some content pointed to from the registry will be available only to students or faculty or staff – and there will be a great deal of content that is available freely to the public.  The outline below begins to express the requirements that seem implicit in much of the discussion that has gone on to date.    

1. GENERAL REQUIREMENTS
1. The Registry will consist of records only-the objects may exist on servers throughout campus
2. The registry will contain records for objects that are freely available to the public
3. The registry will contain records for objects that are available only to students or faculty or staff
4. The registry will be based on concepts established in the “Functional Requirements of Bibliographic Records” documentation which identifies three groups of entities (The following is excerpted from http://www.oclc.org/research/projects/frbr/index.shtm):

                                                               i.      Products of intellectual or artistic endeavor (publications)

1.        a work - distinct intellectual or artistic creation

2.        the expression – the intellectual or artistic realization of a work

3.        the manifestation – the physical embodiment of an expression of a work

4.        the item – a single exemplar of a manifestation

                                                              ii.      Entities responsible for the content of these publications (person or corporate body)

                                                            iii.      Subjects of these intellectual or artistic endeavors (concept, object, event, and place)

5. For content objects that have been designated to be of preservation quality the registry will use the concept of  “preservation” information as described in the “Reference Model for an Open Archival Information System (OAIS)”.

                                                               i.      Preservation Description Information

1.        Provenance information

a.        History of the object

2.        Reference  information

a.        Describes mechanisms used to provide identifiers for the bjects

3.        Context information

a.        Why an object was created  and how it relates to the content information objects

4.        Fixity information

a.        Authentication mechanism to ensure that object has not been altered 

6. The registry should be able to handle the distinction between items that are going to be around for a limited time as well as those that we plan to store for archival purposes.
7. The registry must be able expose its metadata to other applications
8. Must allow for various levels of description.  Some items may simply be a top level description of a web site.  Some may be the third image in the second folder in the fourth box of the fifth collection that combined with 8 other collections makes up the fourth special collection held be the 3rd archive on west side of campus..."granularity?"
9. Support restricted vocabularies
10. Queries must work whether the searcher is using the "academic" or "popular" terms for searching.
11. It must also allow for partial cataloging of items
12. We must be able to deal with the distinction between digital surrogates and born digital objects through metadata.
13. The registry must be able to "ingest" information from a select number of common external data stores (access, MS SQL, Oracle, LDAP????)
2. RECORDS-Records will consist of fields that will adhere to a metadata schema.  This schema will include 3 levels of metadata:  “required”, which is the minimal set of fields necessary to establish the record in the catalog; “item” level records, which are records describing individual digital objects; “collection” level records, what are records referring to a set of objects.
1. Records can “expire”
2. Records may come from any UT department
3. There will be required field and optional fields

                                                               i.      Required fields

1.        Each record must have a URL.

2.        Each record must have an indication of who is responsible for that content being described in the record.

3.        Earch record must have a title.

4.        Each record must have a description

5.        Each record must have  subject keywords assigned.

6.        Each record must have a creation date (for the object being described).

7.        Each record must have format information.

8.        Each record must have a rights field

9.        Each record must have a usage field

10.     Each record must have an intellectual property field

a.        Who does the content belong to

11.     Each record must have an “acknowledgements” field

a.        Specific acknowledgment language

                                                              ii.      Optional fields

1.        Records may have “Learning Resource Type” field.

2.        Records may have a “Learning Context” field. 

3.        A single record may describe individual digital files or aggregations of files.

4.        Record may have a collection field

3. CATALOGING INTERFACE
1. There must be a Web-based cataloging interface
2. There must be an graduated access control system capable of enabling levels of access  to the data; e.g. some users can write but not edit, some can write and edit but not delete, some can write, edit, and delete.
3. There must be a “public” interface that enables an end-user to submit a record with no authorization.
4. Must be able to copy records
5. Must be able to input material in batch mode
6. Must be able to input records item by item
7. Must be able to edit records
8. Must be able to save and delete records
9. Must be able to harvest metadata from external sites.