What is a digital collections platform?

Imagining an Ecosystem:
Selecting a Digital Collections
Platform for the Library
Graham Hukill,
Wayne State University, 2012
What is a digital collection?
What is a digital collections platform?
The lay of the land
Our Model
Lessons Learned and Learning
“In general, creating a work in new media can be
understood as the construction of an interface
to a database.”
-Lev Manowich, “Database as a genre of new media”
“A digital collection consists of digital objects
that are selected and organized to facilitate their
discovery, access, and use. Objects, metadata,
and the user interface together create the user
experience of a collection.”
-NISO, “A Framework of Guidance for Building Good Digital Collections”
“A repository is to storage media as a library is to
shelves. A storage mechanism is needed, but all
public benefit depends upon a repository: a safe
place to put content that also provides
documentation, preservation and access.”
- Richard Wright, BBC
Library Mission
“Wayne State University Libraries advance
scholarship, student learning and faculty
innovation through continuous development
of a library that serves as a national model for
a research University with an urban teaching
and service mission.”
- http://lib.wayne.edu/info/about/#mission
Primary Functions
Store digital objects
Provide access to digital resources
Facilitate search and discovery
Encourage preservation actions and planning
Integrate with Library infrastructure
A steward in the “Information Lifecycle”…
But we’re not going to go down that rabbit hole today…
ADMIRAL Data Management lifecycle diagram:
Digital Collections Ecosystem:
Simple Edition
Digital Collections Platform vs.
Institutional Repository (IR)
• Are they different? Sure!
• Can they be the same? Absolutely!
• Is this confusing? You betcha!
• Our IR, DigitalCommons@Wayne, focuses on preserving and
making accessible scholarly work generated by the University
– DC is very good at what it does
– Will become a modular / contributing component of our
greater Digital Collections “Ecosystem”
Digital Collections Ecosystem:
“From Ingest to Access: A Day in the
Life of a HathiTrust Digital Object”
“Islandlives Workflow”
"A watch, undone": http://www.flickr.com/photos/32943218@N05/6308833700/
Open-Source Platforms
• DSpace
– Originally MIT and HP labs, now Duraspace. open-source, turnkey, IResque, OAIS compliant
• Fedora Commons
– Originally Cornell, now Duraspace, open-source. NOT turnkey, no
discovery layer, VERY customizable
• Omeka
– PHP based, originally designed for metadata and online exhibits
• Eprints
– Early pioneer in digital repositories, IR-esque, emphasis on text
• “the appropriate strategy to avoid the
perceived future deadlock in development is
to abandon local solutions and seek to utilize
other popular Open Source tools that will
more appropriately meet the needs of the
- http://biecoll.ub.uni-bielefeld.de/volltexte/2011/5167/index_en.html
• The best of both worlds in one platform? It
might not be that simple…
Other Key Players
• Apache Solr
– Full-text, faceted search built on Apache Lucene
– Fast, open-source, REST-like interface
– Very popular, “Best of Breed”
• Blacklight
– Discovery Interface for Solr
– Missing link between users and powerful metadata
– Ruby on Rails Gem
Other Key Players
• Hydra:
“Hydra is an ecosystem of components that lets
institutions deploy robust and durable digital repositories
(the body) supporting multiple “heads”: fully-featured
digital asset management applications and tailored
workflows. Its principle platforms are the Fedora
Commons repository software, Solr, Ruby on Rails and
- http://projecthydra.org/
Yale’s “Office of Digital Initiatives and Assets”
“This diagram does not identify every system at Yale. Rather, it illustrates the type and
relationship of modular systems that constitute a comprehensive approach to digital content
- http://odai.yale.edu/planning-and-architecture
"The key to the interoperability of digital cultural
content, and in fact any digital collection,
is consistency . . . inconsistent digital collections
require … more complex, costly and unreliable
systems and processes.”
- http://mirror.dlib.org/dlib/january02/gill/01gill.html
“...one thought to consider is that we are trying
to tackle technology incompatibility which is
often caused by proprietary software and
formats. Therefore, it’s a little ironic to
implement proprietary software tools to
manage this problem.”
Desired Platform Characteristics:
• Open-source, Free
• Flexible, Iterative, and Agile
• Modular
• Support Preservation Activities
• Facilitate Discovery
• Robust and API-like methods
for Access
Our Current Infrastructure
Mirrored Linux / Ubuntu file system
Discovery through catalog
Ad hoc collection interfaces
– not necessarily a bad thing…
Current Infrastructure
Ideal Infrastructure
The Reality
• When envisioning
different layers,
components are
• Researching different
approaches and
software reveals new
pieces and eliminates
Our Model
A work in progress…
“…a minimal set of
requirements and
services that must be in
place to effect the
infrastructure of a
universal, open, widearea digital information
infrastructure system
(“the System”).
- http://dx.doi.org/10.1007/s00799-005-0128-x
A HathiTrust subset
File Structure
• Simple Linux File System at first, transition into OAIS inspired
system (Fedora, DSpace)
• Fully mirrored storage
• Self-describing File System
– METS files will be duplicated in SQL storage and in file
METS objects
– Descriptive and Structural information about objects
– wrapper for other metadata formats (DC, MIX, MARC, PREMIS, etc.)
Central to the overall architecture
Flexibility: constituent pieces can be located anywhere
Consistency: normalized metadata for all objects
Interoperability: digital objects accessible by various systems
Widely adopted
<file ID="FID1" GROUPID="2" USE="Image-Service-MedRes">
<FLocat LOCTYPE="URL" xlink:type="simple" xlink:href="letter_city00001.jpg"/>
<div ORDER="1" LABEL="page01" DMDID="dm2" TYPE="page">
<fptr FILEID="FID1"/>
METS objects
Provides a framework to handle
different kinds of digital objects.
#1 – somewhat typical “digital
#2 and #3 – eTexts
#4 – “atomistic” digital object,
containing select pieces of other
Similar to “collection” at scale
and breadth perhaps, but
fundamentally a bit different
Metadata Index & Database
Searchable store of METS files
Backend to Discovery layer (Blacklight)
Ancillary to Access layer (various components)
– Java “servlet”, HTTP interface
– Full-text search, faceted search, database integration
– JSON, XML, PHP, Ruby, Python, XSLT, Velocity and custom Java binary
output formats
– Industry standard: used by Internet Archive, HathiTrust, The Guardian,
Instagram, you get the picture…
Discovery Layer
• Critical component, often first point of contact with
collections for users
• Search within, and across, collections
• Search all material types (visual, text, A/V)
• Much of the “look and feel” of a digital collection and the
digital collection environment
• Potential for accessing and presenting digital objects as well
Current landscape...
- Howard Besser, “The Next Stage: Moving from Isolated Digital Collections to Interoperable Digital Libraries.”
What it could look like...
- Howard Besser, “The Next Stage: Moving from Isolated Digital Collections to Interoperable Digital Libraries.”
Discovery: Blacklight
NC State – “Historical State 125”
Discovery: Blacklight
University of Wisconsin Madison – Library Catalog
Discovery: Blacklight
Stanford – “SearchWorks” Library Catalog
Discovery: Blacklight
New York Public Library
Andre Studios 1930 – 1941
Receives requests from discovery layer
Read appropriate METS file
Pull digital objects from “storage”
Provide access / interface for these materials
• Types
– Images
– eTexts
– Audio / Video
Access – eText Reader
Not so fast…
METS files
• Creation of METS files can
be time consuming, and
very different for different
kinds of materials
• At right, simplistic model
of how to ingest and
create METS files for eTexts
• “METS creation tool”,
while handy, requires
human intervention
Not so fast…
METS files
• Archivematica
– Batch, recursive processing of items
– Focused on preservation (AIP creation)
– very descriptive METS files (JHOVE, Droid, etc.)
• Archivist's Toolkit (AT)
– Java application, difficult to batch process
– Requires SQL backend, finicky
We’re on the right track!
• Asking fundamental questions about our digital materials,
what digital collections are, has been key
– Sometimes by envisioning an ideal solution, we uncovered others who
had trod similar ground
• “These “Best of Breed” solutions should always be considered
over local homegrown tooling.”
- http://biecoll.ub.uni-bielefeld.de/volltexte/2011/5167/index_en.html
– Utilize the hard and spectacular work of others
– But sometimes the works has not been done. If not, why not?
• In these diagrams, sometimes the “arrows” are the most
challenging bits to reconcile
– API interactions
– Often require crosswalking or transformation of
• Internet Browsers are the most powerful and consistently
updated pieces of software around
– Internet Archive’s Book Reader pure HTML, CSS, and Javascript
• Discoverability and Access across collections should be
available from one location
– storage and metadata layers can provide collections and
objects through API’s to other – and even yet unforeseen systems
• Occasionally step back and ask:
– Is this system more consistent and versatile than before?
– Will it scale? can it evolve?
– Is this what our users want? what our collections need?
Graham Hukill,
Wayne State University, 2012
ADMIRAL Data Management lifecycle diagram:
Stanford "SearchWorks" screenshot: http://searchworks.stanford.edu/
NC State screenshot: http://historicalstate.lib.ncsu.edu/
John Hopkins University Youtube screenshot: https://catalyst.library.jhu.edu/
UW Madison screenshot: http://search.library.wisconsin.edu/
NYPL screenshot: http://andrestudios.nypl.org/browse
"A watch, undone": http://www.flickr.com/photos/32943218@N05/6308833700/
"Blueprint": http://www.flickr.com/photos/wscullin/3770016707/
"Donovan School (Ann Arbor, Mich.) Miss Lily E. Goodhew teacher, Feb. 1911.": http://quod.lib.umich.edu/b/bhl/x-bl000298/bl000298

similar documents