Public Access to Fully Digitized Collections at the

Public Access to Fully Digitized Collections at the Archives of American Art NARA Preservation Conference Digitizing for Preservation and Access: Past is Prologue March 26, 2009 Karen B. Weiss weissk@si. edu

Overview of Presentation n Introduction to Collections Online n Background on the Archives’ approach to public access and to large-scale digitization n Building the technical infrastructure n Workflows n Enhancements and next steps

Collections Online: Large Scale Digitization at the Archives of American Art http: //www. aaa. si. edu/collectionsonline n Entire collection vs. selected items n Currently, 67 collections; 450 linear ft. ; 600, 000 digital files n Built upon fundamental archival approaches

Folder titles provide links to images

Image Viewer Provides Access to Contents in Each Folder

About the Archives of American Art n Founded in 1954 in Detroit n Original goal to locate and microfilm for widespread n n access to primary resources on American art; not to collect and preserve originals Immediately offered donations of papers and collecting and preservation became important Became part of the Smithsonian in 1970; AAA currently holds approx. 6, 000 collections 15, 000 linear ft; 2, 000 oral history interviews http: //www. aaa. si. edu

Collections Online: Basic Approach n All descriptive metadata derived from the EAD finding n n aid Access to the digitized documents is through folder level access instead of item level access Box and folder listing forms the file structure for the scanning technician to save the digital files, as well as the primary descriptive metadata for discovery of the digital files. In house Digital Collections Database’s Finding Aid Table contains the XML data; dynamically driven web presentation + image viewer Internal web based workflow support

Terra Foundation For American Art Digitization Grant n 2005 – Archives of American Art was the first grantee of the Terra Foundation for American Art. Terra’s mission: To help create a worldwide community for American art presentation, research, and interpretation, the foundation is dedicated to facilitating partnerships and communication and to increasing accessibility to needed scholarly resources. http: //www. terraamericanart. org

Terra Grant Proposal Requested funding to: n Replace microfilming n Support team of processing archivists, digital imaging technicians, programmer/web developer n Purchase scanning equipment n Digitize and provide access to 105 collections/ 500 linear ft. of most significant collections, some from microfilm n Selection approved by Terra; represents a cross section of significant collections; minimal contemporary content n http: //aaa. si. edu/collections/terra_collections_list. cfm #list

Terra Digitization Grant = Processing Grant Collection level digitization requires collections be fully processed and EAD finding aid prior to digitization, but… n Very few finding aids until early 1990 s n Used microfilm labels and catalog cards as main source of description and in lieu of finding aids n Full retrospective MARC cataloging 19851988; laid the collection level foundation for implementing finding aids

Microfilm: The Prequel n In-house microfilm operation n Microfilmed for access n Opened network of regional centers and adapted traditional interlibrary loan n Gained understanding of production; able to microfilm approximately 30% of holdings

50+ Years of Microfilming Supports. Large-Scale Digitization Microfilming with Zeutschel microfilm camera 2002 Scanning with Zeutschel 10000 AI Scanner 2005

Building Blocks – Basic Principles n The focus is on access to collections (which differentiates our approach from archival management systems) n The information resides in a database, and that drives the online presentation, as opposed to using a stylesheet transformation n Rich, descriptive, and contextual metadata can be derived from the structured and tagged data found in EAD finding aids.

Building - Technical Overview n All programming based on Adobe Cold. Fusion (with some Java programming) n All data is stored in a MS SQL Server database – the same database used to store our existing collection and digitization information n The application is built using an objectoriented approach that organizes the structure of the system based on function (i. e. EAD ingestion, PDF creation, image processing, deployment, etc)

Building the Database (I) EAD/Finding Aid n Designed the database around the archival information, not the EAD encoding

Building – EAD Uploading n Each EAD XML file is passed through a parser that transforms the XML EAD data into an EAD Document Object n Each portion of the EAD object is transformed into its “real world” equivalent (i. e. descriptive information, series, container listing, etc) n Most of the descriptive data for the finding aid gets stored in the Finding Aid table, series information is stored in a Series table, container listing information is stored in the Container table

Building – EAD “Engine”

Building –Workflows for Digitization n Information gathered from the EAD file describes the box / folder hierarchy for scanning n This allows for automatically generating nested file structures for scanning n Scanned images are copied onto Network Attached Storage (NAS) devices before processing

Building the Database-Images n Struggled to figure out how to incorporate links to images in the EAD using <dao>, <daogrp>, etc. n Decided instead to associate images based on container / folder heading data n Ended up being a very flexible system for image association

Building - Access to Images n Image converting, resizing, and watermarking is batch automated on a collection-wide level n Database records for digital images are created at the same time the images are processed n Mistakes happen – the workflow supports refreshing metadata for images without processing, or rescanning only a portion of a collection without having to reprocess all of the images

Automated Workflows n Initial uploading of the archivists’ XML EAD finding aid into n n n n finding aid repository; subsequent check-in and check-out Creation of an internal Collection Online “stub” site for the collection Generation of the directory structure (based on the EAD XML container listing) for the digital imaging technician to use when scanning Batch processing of the TIFF files to create three jpg derivatives and watermarking Linking a representative image to each series description for a visually appealing presentation Linking of the digital files to the appropriate folder titles Generation of a pdf finding aid accessible from the collection site Deployment to the final presentation on the web Versioning support

Reporting Capabilities

Web 2. 0 Enhancements n Comments n Volunteers for transcribing letters n Links to related Smithsonian collections n Links to related exhibitions, lectures, events

Website presentation n New awareness of importance of website presentation and interface design n New Webmaster position n Constantly analyzing use and feedback n Implemented usability techniques and conducted usability studies

Feedback and improvements Since launched in November 2006: n Replaced search engine n More intuitive, user-friendly design n Image Viewer improved n Printing option installed n New stylesheet for non-digitized collections

Next Steps n Better integrate digital and non-digital collections information n Integrate fully digitized collections with itemlevel /selected images and other resources n Utilize EAD for subject, date, and controlled name access n Integrate digitized audio/video n Transcriptions n Explore open-source platform