Survey of Music Materials utilizing

Encoded Archival Description (EAD)

Prepared for the

International Music Metadata Projects Working Group

 

by Lois Schultz

January, 2001

 

 

Survey of Music Materials utilizing Encoded Archival Description (EAD)

 

This information has been based on a small sample of the hundreds of music finding aids that are available on the Web.  For a discussion of finding aids, see the separate description.  Music materials have, quite naturally, been included in these collections and archivists, many of them music specialists, have created finding aids for them.  The collections range from Jewish songsters, Benny Goodman papers, Rosa Ponselle papers, Deems Taylor papers, Charles Griffes papers, scrapbooks; popular sheet music; big band charts, sound recording archives, film archives, etc.  To give some indication of the scope of use, the University of California, Berkeley has over 100 finding aids for music, the University of California, Los Angeles has 28, and the New York Public Library has 46.  All three libraries are creating more of them. EAD has been heartily embraced by the archival community and they have been not only creating new finding aids in EAD, but converting printed documents at an astonishing rate. 

 

Implications for Music Materials

The music community has spent considerable effort in developing guidelines for the inclusion of appropriate description and access in online catalogs.  We have been somewhat less active in the archival community, with the result that, where music materials are encountered in repositories, the finding aids have generally not been specifically designed for music retrieval.  For example, the universities at Berkeley and Los Angeles have an impressive number of finding aids available for use, but the SGML tagging has been adapted from the standard and doesn't actually allow the kind of specialized retrieval that is permitted in MARC records (e.g., the use of qualifiers/tags for composers and performers).  Some of the members of our community have been very active in cataloging and arranging these collections (Stephen Davison of UCLA and Judy Tsou, formerly of UC-Berkeley), but they have been using the standard DTD for description and access rather than one that is specially designed for use with music.

 

Study of Specific Sites

Methodology

For each site I looked at more than one finding aid, but selected a typical one to study more intently.  I also looked more closely at finding aids or sites that had been specially adapted for music.  Selections of these documents with more specific information about the display and tagging of the finding aids are given later.  All of the sites using EAD follow the standard DTD and are fairly uniform in their presentation and indexing.  Only the Historic American Sheet Music project, in the Duke implementation of it, uses single item records in a database that generates searches and pages "on the fly."  This scheme was adopted in order to provide very specialized access to a large number of files (over 3,000).  A few of the sites (listed at the end) were actually written in HTML, so I didn't study them extensively.

 

Duke University

Project: Duke University

Sonny Burke Collection

I.          Project Description

Bibliographic documentation: EAD

Database structure: Hierarchical

Metadata

Administrative: scanned images, text indexing

Structural: EAD

DOCTYPE ead PUBLIC "-//Society of American Archivists//DTD ead.dtd (Encoded Archival Description (EAD) Version 1.0)//EN" "./dtds/ead.dtd

Level of detail: item level

II.         Background documentation:

This finding aid is designed to provide item level access to the big band "books" in the collection.  There are no image or sound files integrated into the document.  In style, it is very much like the other finding aids at Duke, with slight modifications in the "labels" so that more specific searching is permitted.

III.       Information retrieval

Searching is as specific as standard EAD allows: e.g., personal names, dates, title.  While specific searching by arranger or copyist isn't directly available, the use of labels may allow a certain level of proximity searching.

            Client-server architecture: Yes

            Supports Boolean searching: Yes

            Supports browsing: Yes

IV.       Accessibility/display:

            SGML/XML compliant: Yes (XML)

            SGML browser (e.g., Dynaweb/Panorama): Yes

            HTML display  : Yes

            Sound files: None

            Image files: None

 

Project: Duke University

Historic American Sheet Music (as available at Duke, see American Memory at LC for another version)

I.          Project Description

            Bibliographic documentation: EAD

            Database structure: Hierarchical

            Metadata

                        Administrative: scanned images, text indexing

                        Structural: EAD

            Level of Detail: Individual titles, composers, lyricists, illustrators, engravers, arrangers, publishers, first line and refrain of text, dedicatee, Library of Congress subject headings, Library of Congress thesarus for graphic materials terms, instrumentation, and special "subject" headings designed for the project.

II.         Background documentation:

The project provides access to over 3,000 pieces of sheet music.  It functions as a hierarchical database rather than a standard EAD finding aid of SGML tagged text.  Indexing was done by Lois Schultz and the database structure, finding aid/display was designed and implemented by Stephen D. Miller.

 


III.       Information retrieval

This finding aid is designed to provide a very high level of access to the musical, artistic, and social content of the sheet music published in the U.S. between 1850 and 1920.  The searching is enhanced through specialized retrieval based on the tagged data.

            Client-server architecture: Yes

            Supports Boolean searching: Yes

            Supports browsing: Yes

IV.       Accessibility/display

            SGML/XML compliant: Yes

            SGML browser (e.g., Dynaweb/Panorama): Yes

            HTML display  : Yes

            Sound files: None

            Image files: JPEG

            Other: permits "page turning" so that you may move directly through the music.

 

Project: Harvard. Loeb Music Library

Based primarily on the Eubie Blake finding aid

I.          Project Description

            Bibliographic documentation: EAD

            Database structure: Hierarchical

            Metadata

                        Administrative: text with indexing

                        Structural: EAD

II.         Background documentation:

The collection consists mainly of materials by and about Eubie Blake, primarily from the 1970's, collected by William D. and Peggy Smith.  Finding aid by Sarah Adams

III.       Information retrieval

            Client-server architecture: Yes

            Supports Boolean searching: Yes

            Supports browsing: Yes

IV.       Accessibility/display

            SGML/XML compliant: Yes

            SGML browser (e.g., Dynaweb/Panorama): Yes

            HTML display  : Yes

            Sound files: None

            Image files: None

Level of detail: item level, based on standard search terms

 

Project:NYPL finding aids

Description based on: Charles Griffes collection

I was unable to view the SGML tags of the NYPL finding aids, so the information is incomplete

 

I.          Project Description

            Bibliographic documentation: EAD

            Database structure: Hierarchical


            Metadata

                        Administrative: text with indexing

                        Structural: EAD

            Level of detail: item level, based on standard search terms?

II.         Background documentation: The collection consists of mostly music manuscripts of American composer Charles Tomlinson Griffes.  This finding aid also includes several manuscripts that were cataloged prior to the accession of the main part of the Griffes collection (classmark JPB 84-1).  Donated by Mrs. A. Marguerite Griffes, the composer's sister, in 1954.

III.       Information retrieval

            Client-server architecture: Yes

            Supports Boolean searching: Yes

            Supports browsing: Yes

IV.       Accessibility/display

            SGML/XML compliant: Yes

            SGML browser (e.g., Dynaweb/Panorama): Yes

            HTML display  : Yes

            Sound files: None

            Image files: None

Level of detail: item level, based on standard search terms

 

Project: University of California, Berkeley

Description based on: 1750 Arch Tape Collection and Bay Area Musicians Collection

I.          Project Description

            Bibliographic documentation: EAD

            Database structure: Hierarchical

            Metadata

                        Administrative: text with indexing

                        Structural: EAD

            Level of detail: box level with some information about the contents, standard search terms

II.         Background documentation: provides access to a large collection of sound recordings and video cassettes; collection is arranged by date, with access to information about each concert.

III.       Information retrieval

            Client-server architecture: Yes

            Supports Boolean searching: Yes

            Supports browsing: Yes

IV.       Accessibility/display

            SGML/XML compliant: Yes

            SGML browser (e.g., Dynaweb/Panorama): Yes

            HTML display  : Yes

            Sound files: None

            Image files: None

Level of detail: item level, based on standard search terms

 


Project: UCLA Music Special Collections

Description based on: CBS Television and Film music collection

I.          Project Description

            Bibliographic documentation: EAD

            Database structure: Hierarchical

            Metadata

                        Administrative: text with indexing

                        Structural: EAD

            Level of detail: item level, based on standard search terms

II.         Background documentation: A large collection of music for most of the television and several motion picture productions of the CBS/MTM (Mary Tyler Moore) studios in Burbank.  This on-line finding aid is an unedited version of the original finding aid.

III.       Information retrieval

            Client-server architecture: Yes

            Supports Boolean searching: Yes

            Supports browsing: Yes

IV.       Accessibility/display

            SGML/XML compliant: Yes

            SGML browser (e.g., Dynaweb/Panorama): Yes

            HTML display  : Yes

            Sound files: None

            Image files: None

Level of detail: item level, based on standard search terms

 

Projects that consist of HTML pages only or lack music content

 

Project: Margaret and Franklin Steele sheet music collection (HTML finding aid only)

I.          Project Description

            Bibliographic documentation: EAD

            Database structure: HTML pages

            Metadata: HTML, no metadata used for indexing purposes

 

Project: Duke University

America Votes: Presidential Campaign Memorabilia

I.          Project Description: Online exhibit that includes some music

            Bibliographic documentation: HTML

            Database structure:

                        Other (e.g., not a database): HTML pages

            Metadata: HTML, no metadata used for indexing purposes

 

Project:American Heritage Project (http://sunsite.berkeley.edu/amher/)

I.          Project Description

This was a small test project and doesn't include any music materials

 

 

 

 

Description of Finding Aids

 

Based on information at the Duke Rare Book, Manuscript, and Special Collections web site: http://scriptorium.lib.duke.edu/findaids/findaids_tech.html

 

Finding Aids are descriptive tools that serve as the primary point of intellectual access to archival collections.  A finding aid may take on various shapes and be called by a variety of names (e.g., finding guide, inventory, register) but generally performs a dual function -- it serves both as the primary access tool used in archival reference of a collection, and as a complex management device used for the administration of that collection.  Prior to the late 1980's, most of the description of manuscripts was done on catalog cards or in printed books (often reproductions or transcriptions of the catalog cards)

 

Finding Aid Structure

 

These finding aids typically consist of several elements.  They include a title page which gives the formal title and dates of the collection, along with acquisition and restrictions information, and a statement of extent (measured in both linear feet and estimated item count).  The next element is a biographical/historical summary of the significant events in either the life of the person or the corporate body or organization whose papers/records constitute the collection.  Following this is a descriptive summary of the scope and contents of the collection, including information on the overall organization of the collection, materials of particular significance, and important topics and/or persons represented.  The detailed contents of each collection are given in the container list that is arranged according to the actual physical organization of the materials.  The content and scope of each organizational unit in the collection (typically called a "series," with units within series known as "sub-series") are described in a brief note at the beginning of each series within the container list.  This note is followed by a box list that gives box numbers and folder titles within each box.  Some earlier finding aids have a separate "description of series" in which the individual organizational units are described separately from the actual container list.  Current practice is to combine such descriptions with the corresponding section of the container list.  Additional elements that may or may not be present in finding aids include bibliographies of published writings, processing notes, and special indexes.

 

 

Encoded Archival Description

Description

 

Encoded Archival Description (EAD) serves archives and manuscripts repositories and was begun by Daniel Pitti and developed by the Network Development and MARC Standards Office of the Library of Congress with the Society of American Archivists.  The standard, which became available in Sept., 1998, may be retrieved from the LC web pages: http://lcweb.loc.gov/ead/.  It is used to display and index electronic finding aids and registers for manuscript and archival collections.  The SGML/XML tags mark specific areas of the finding aid (e.g., scope and contents notes, series, description, container lists).  The markup may be highly complex or very simple.  Possible embedded tags include names, dates, titles, etc.  Controlled vocabulary, thesauri, and authority files may be used in conjunction with the standard.  Locally developed files are also supported.  To date, there is no official standard (i.e., a DTD or Document Type Definitions) for music materials within EAD.

 

EAD provides for both collection-level, hierarchically structured finding aids and MARC-like individual cataloging records (Single Item Metadata or SIM) for unique items in the collections.  The SIM records may be linked to the appropriate folder in the finding aid.  The structure provides navigational ability to move easily from finding aid to digital object (e.g., scanned image or sound file) or digital object to digital object within the finding aid, much as individual items in a folder are browsed by researchers using the physical documents in an archive.  Relational database management systems can parse finding aids and related SIM records to provide sophisticated retrieval and navigation.  The development of DTD's specific to a project or discipline allows great flexibility in the use of the standard.  Getting mileage out of metadata includes some sample records.

 

Searching and Retrieval of Information in Finding Aids

 

Searching EAD finding aids is extremely powerful because you may search all or some of the documents in a repository or consortium at the same time.  An example of a consortium using EAD is found in the California Digital Library (http://www.oac.cdlib.org/dynaweb/ead/), which includes finding aids for about 59 libraries.  Because you may specify a search for words near each other (proximity), a much higher level of specificity is achieved in the set of records returned.  While there are still false "hits," the records that include words in close proximity are given a higher rating, and display at the top of the list.  Additionally, you may specify searches as personal names, corporate names, titles, dates, etc.  This means that you have the search capability of a database (e.g., one as structured as MARC) with the flexibility of text.

 

Role of Standard Generalized Markup Language in EAD

Standard Generalized Markup Language (SGML) and its subset, eXtensible Markup Language (XML) are important tools for metadata creation.  SGML uses Document Type Definitions (DTD) as standards for different types of material.  DTD's may be developed by user communities or individuals. The markup is not set just to display text and hyperlinks, as in HTML, but allows more specific retrieval.  Parts of the document are specifically labeled for descriptive and access functions.  There are two DTD's frequently used in libraries: Text Encoding Initiative (TEI) and Encoded Archival Description (EAD).  TEI, which is not further described here, is primarily used for marking up literary documents and other texts.  Key projects include the Early American Fiction Project at the University of Virginia (http://etext.lib.virginia.edu/eaf/pubbrowse/) and the Making of America Project at the University of Michigan (http://moa.umdl.umich.edu/) and Cornell University (http://library5.library.cornell.edu/moa/).

 

SGML is a document-creation language that can embed descriptive elements within the document itself.  Parts of the document are specifically labeled/categorized and the DTD's can provide not only such formatting tags as paragraphs, lists, and page breaks, but may also provide access to historical background, ownership, provenance, and controlled vocabulary headings (such as subject headings and genre terms).  They can be as specific as corporate/personal author, title/uniform title, etc.  Markup of the text ensures that the indexing is embedded within the text, so that it will always remain closely linked.  Metadata embedded in the text may be the most reliable method for providing persistent indexing for electronic texts.  The structure of SGML also provides a scheme to describe the original source material (e.g., a manuscript or imprint).  The creation of metadata centers on establishing practices for encoding areas in the document to create fielded metadata within the body of the document, as well as adding description to identify the document itself.  Cataloging records may also exist independently of the electronic source.  EAD is an example of this.  Most libraries have a catalog record in their online catalogs representing the entire collection and a link to the electronic finding aid.

 

SGML is highly extensible, providing for a wide range of electronic documents, including scores and sound recordings.  Typically, user groups, such as the Society of American Archivists, have developed DTDs that specifically fill their needs, so members Music Library Association could develop standards for those materials.  Database management systems are often used in practical applications of SGML to provide sophisticated searching tools and create the record and report the results in SGML or HTML format, depending on the browser used.  Robust database management systems offer efficiencies for management, retrieval, and manipulation.  Extensible Markup Language (XML), a subset of SGML, provides increased flexibility and functionality in hyperlinking and is rapidly becoming the dominant standard.

 

Example of a non-music EAD finding aid

Most finding aids at Duke University are currently using the standard implementation of EAD. The descriptive summary, biography, scope and content notes all follow the standard SAA DTD. In Ex. 1 the access points are specifically marked as "persname," "corpname," etc. within the body of the text.  This type of markup allows searching for words that are relatively close to each other, hence the search "1978 and doris and duke and publisher's and contract" produced 9 hits in all of Duke's finding aids and all were in the same collection of papers.  Thus the ability to use proximity to reduce the number of false retrievals is an important one.  As you might imagine, without such specificity, just the name "Duke" produces a very large retrieval set (5420).  This example demonstrates the power of searching text using appropriate search engines and SGML markup. 

 


Ex. 1.   Standard archival finding aid (excerpt from the Angier Biddle Duke papers)

SGML markup

<did><container type="box">1</container>

<unittitle id="inte">Legal Papers, <unitdate type="inclusive">ca. 1970s-1980s. </unitdate></unittitle>

<physdesc><extent>(3 folders)</extent></physdesc>

<scopecontent>

<p>Publisher's contract related to <persname>Doris Duke </persname>biography, document related to the <corpname>Doris Duke Trust </corpname><date>(1978),</date> and documents from a<unitdate type="inclusive"> 1978 </unitdate>lawsuit involving <persname>"Pony" Duke.</persname></p>

</scopecontent></did>

 

Displays as:

Box 1

Legal Papers, ca. 1970s-1980s. (3 folders)

Publisher's contract related to Doris Duke biography, document related to the Doris Duke Trust (1978), and documents from a 1978 lawsuit involving "Pony" Duke.

 

Bibliography

Jean Hudgins, Grace Agnew, Elizabeth Brown, Getting Mileage out of Metadata: Applications for the Library. Chicago: American Library Association, 1999 (LITA guides #5).

 

IFLA. Study Group on the Functional Requirements for Bibliographic Records.  Functional Requirements for Bibliographic Records.  Munich: Saur, 1998.  http://www.ifla.org/VII/s13/frbr/frbr.pdf

 

Janice E. Ruth, "Encoded Archival Description: A Structural Overview," American Archivist 60 (Summer 1997).

 

Webliography for EAD

 

EAD standard

http://lcweb.loc.gov/ead/

http://jefferson.village.virginia.edu/ead/cookbookhelp.html

Early English Fiction

University of Virginia (http://etext.lib.virginia.edu/eaf/pubbrowse/)

Making of America Project

University of Michigan (http://moa.umdl.umich.edu/)

Cornell University (http://library5.library.cornell.edu/moa/)

Richard Rinehart, Museums and the Online Archive of California (http://www.bampfa.berkeley.edu/moac/standards/index.html) This includes a history of how EAD became part of the Online Archive.

 

 

 Appendices