METS Implementations at Harvard University Library

Implementation 1:  Asynchronous delivery of biomedical image stacks

Description:  The Biomedical Image Library is a publicly accessible
retrieval system for original digital micrographs that have been produced
in support of basic biological research. XML descriptive metadata is held
in a catalog, while METS files that identify the components of each "image
stack" and the original images (TIFFs) and associated administrative and
technical metadata reside in Harvard's Digital Repository. METS files for
each stack simply list the components; relationships among files are
expressed in descriptive metadata (series type, file naming convention,
step type, number of steps, channel list, channel unit).

Records in the catalog represent a project, which may contain hundreds of
image stacks created with various samples, preparations, and instrument
settings. For each stack, the catalog displays the thumbnail of one
representative image from an image stack. The thumbnail links to both a
full size version of the single image for online viewing and to the
downloadable image stack. Users can select one or more image stacks from
the project, and when they export their saved set, the metadata and the
image files are zipped for pickup and the user is notified. The zip file
unpacks into a file structure that preserves the hierarchy of the
original project(s).

Implementation Date: In production as of March 2003

URL to production system:
http://nrs.harvard.edu/urn-3:hul.eresource:bioimlib

URL to documentation:
Descriptive schema: http://hul.harvard.edu/ois/xml/xsd/bil/bil_schema.xsd

Asynchronous delivery documentation: coming soon.

Availability of tools:
        Written to integrate with Harvard Digital Repository, so not
generalized for external use.

Contact: Lee Mandell, Programmer/Analyst, Office for Information Systems,
[log in to unmask]


Implementation 2:  Preservation Audio

Description: Using METS to associate archival master, production master,
and deliverable audio files, along with associated technical and process
history metadata, AES-31 audio decision lists, opaque vendor-specific
processing files, and waveform reduction files. AES-draft Core audio will
be extracted from AIFF files. This file is intended for use by audio
preservation engineers. It is not being used to drive delivery to
end-users. Locally-written schema-driven XML editor used to capture
processing history. METS toolkit used to construct METS and to aggregate
package for deposit into Digital repository.

Implementation Date:  Summer 2003

URL to production system: N/A

URL to documentation:  Not yet.

Availability of tools:  METS toolkit available now. Schema-driven XML
editor will be made available.

Contact: Stephen Abrahms, Digital Library Program Manager,
[log in to unmask]


Implementation 3: Page-turned objects

Description: Convert current MOA2 page-based applications to METS.
Currently, the Page-Turner itself uses MOA2, and related and auxiliary
systems such as a complex-object maintenance system, the Digital Imaging
Lab's production systems, the library's full-text indexing system, and the
Digital Repository deposit system are designed to interact with MOA2
files.

Implementation Date:  TBD

URL to Production System: Not yet available for new system.

URL to documentation/specs: Not yet available for new system
Availability of tools: TBD

Contact: Stephen Abrahms, Digital Library Program Manager,
[log in to unmask]