Adam Wead
Penn State University
Stranger in a Strange Land
EAD in a MARC-only World
Data Transformation
MARC
245 0 4 a| The 1997 A & E entertainment almanac :
b| an information please almanac /
c| Robert Moses, Alicia Potter and Beth Rowen, editors
MODS
The
1997 A & E entertainment almanac
an information please almanac
Robert Moses, Alicia Potter and Beth Rowen, editors
DC
The 1997 A & E entertainment almanac :
an information please almanac /
Robert Moses, Alicia Potter and Beth Rowen, editors
Fetish metadata
- PBCore
- VRA Core
- Premis
- METS
- EAD
A Tale of Two Silos
- MARC data via Millennium
- archival collections in Archivists' Toolkit
- digital content in collections
EAD to MARC
- very common, good solution, but...
- granularity of discovery
- we wanted items within collections to be visible
- we needed the finding aid displayed
librarian is to the MARC record,
as the archivist is to the finding aid.
Finding Aids
- big documents
- anywhere from a few to thousands of items
Transformation to EAD
- added markup to the exiting format like HTML
- retained its monolithic structure
- remained order-dependent
Our Unique Collections
- significant items where in collections
- some collections had thousands of items
- how where you going to find that?
Transformational Technologies
Tech Components
- presentation: html
- searching: Blacklight
- indexing and storage: Solr
- Ruby, Rails
Transforming EAD
Breaking up is hard to do...
- one piece for the finding aid
- one piece each for every series, sub-series, folder, etc.
- one piece each for every item
- thousands of pieces per collection
Searching and Relevance
- MARC-based items interfiled with items from collections
- boosts the relevance of items, not just collections
- exposes items across and within collections
- faceted search capabilities
- minimal Schema.org integration
- container-level pieces discarded from search results
Presentation
- pieces are re-assembled for finding aid display
- only relevant sections are displayed
- default view does not show the entire finding aid
- because people don't read web pages [1]
Digital Content
- video was highest priority
- digitized all induction footage, other important series
- decided on PBCore as our metadata standard
Deeper Technology Stack
- Hydra
- adds Fedora repository for storage and preservation
- enabled description and arrangement of video
- PBCore records attached to pieces of an EAD
Did this work?
- yes, but sustainability is unknown
- requires a full-time developer
- unique approach to EAD indexing that doesn't fit other institutions
- self-contained, closed system
- dependent on specific technologies
What does the future hold?
- linked data
- MARC has BIBFRAME
- EAD has no Bibframe equivalent, yet
- most other schemas do not either... MODS does, pbcore does not
- TOO MANY SCHEMAS!
Next Steps
- ontologies for our silos
- ArcLight
- Linked Data Platform (LDP)
- ArchivesSpace
- implementing linked data
Adam Wead
Penn State University
Notes
- http://uxmyths.com/post/647473628/myth-people-read-on-the-web