I thought that subscribers to the METS and PIG lists might be interested
in some of the features of the new PREMIS schema that were not discussed
in the more general announcement of PREMIS 2.0 that went out yesterday.
1. We have combined what was 5 separate schemas (PREMIS container, Object,
Event, Rights, Agent) into one. This will facilitate maintenance and will
have no practical effect on how it is used. There are 5 root elements
defined, premis, object, event, rights, agent and <premis> may be used as
a container or not. This means that in the case of implementing in METS,
any of the elements may be used as root in appropriate metadata sections
of METS (as was possible in the previous version).
2. We have implemented most of the suggestions made by LANL early in the
PREMIS review process. This includes defining all elements globally so
that they can be reused alone if desired. Secondly, we have separate
definitions for the three levels of object: representation, file, and
bitstream. This will allow us to validate the applicability of a given
element for a specific level of object, which we were unable to do before.
3. We are including a mechanism for extensibility for a number of
elements. This is documented in the data dictionary as well. Thus,
extension elements may use structures from another XML namespace. The
elements that allow for extensibility are:
creatingApplication
environment
eventOutcomeDetail
keyInformation
objectCharacteristics
signatureInformation
significantProperties
rights
Note that the objectCharacteristicsExtension is intended to include format
specific metadata from other schemas-- to be used if an implementer wants
to keep all technical metadata together rather than separated as in METS
in separate techMD sections.
4. We have incorporated some additional constructs for expressing date and
date/time. These are in the draft PREMIS schema now, although we may issue
it as a separate schema later on so that other XML schemas may use it
(e.g. MODS). This includes the ability to encode a date using ISO 8601
basic format (i.e. without the hyphens, which is not covered by xs:date,
but is covered by ISO 8601) and the ability to express special types of
dates. This is like a profile of ISO 8601, discarding unnecessary features
and extending it with features that are needed. Some examples of special
dates that are included are questionable dates, open dates, approximate
dates. For now we are calling this "Extended Date/Time format: edtf"
5. In the next incremental version we will include a mechanism to be able
to specify controlled vocabularies used with PREMIS and validate
them. This will be based on work we are doing establishing a database to
maintain PREMIS controlled values. Those that are in the data dictionary
as "suggested values" will be included in the database and will be able to
be validated by an XML schema. The scheme (or authority) for the
controlled vocabulary will be obvious from the URI used for the controlled
value. Since this is a work in progress, it is not yet ready, but these
are the plans. We will encourage other implementers to contribute their
controlled vocabularies to the database.
These are the key features. There are also numerous other changes,
including of course all changes to the data dictionary that were made in
version 2.0. Because of the restructuring and significant changes we made,
we have released it as a draft schema so that we can have a one month
comment period.
Please take a look and send comments by April 25, 2008 to Ray Denenberg
([log in to unmask]). Or discuss on this list.
Rebecca
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
^^ Rebecca S. Guenther ^^
^^ Senior Networking and Standards Specialist ^^
^^ Network Development and MARC Standards Office ^^
^^ 1st and Independence Ave. SE ^^
^^ Library of Congress ^^
^^ Washington, DC 20540-4402 ^^
^^ (202) 707-5092 (voice) (202) 707-0115 (FAX) ^^
^^ [log in to unmask] ^^
^^ ^^
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
|