EAD Application Guidelines for Version 1.0


Chapter 3. Creating Finding Aids in EAD: Continued


3.5.2. Describing the "Parts": Nested Components

Once an archivist has finished describing in a general way the "whole" of a body of records or papers, the focus typically shifts to describing one or more of the collection's parts, as explained in the discussion of multilevel description in section 3.4. In EAD, these parts are known as Components <c>. In the following sections, we explore the nature of components and address such questions as these:

To those archivists accustomed to building hierarchical finding aids, especially those familiar with ISAD(G) and RAD, the answers to these questions will be familiar. In the spirit of ISAD(G), component descriptions in EAD inherit information from both the description of the whole and from any preceding hierarchical component levels. At every level, the descriptions can utilize the same essential data elements.

3.5.2.1. What Is a Component? <c>

A component may be an easily recognizable archival entity such as a series, subseries, file,(63) or item, or, as the examples below illustrate, it may simply be any level or stage within the descriptive hierarchy. Components not only are nested under the <archdesc> element, they usually are nested inside one another. For example, archivists would agree that series are parts or components of collections, fonds, or record groups. Similarly, subseries are parts not only of their parent series but of their "grandparent" unit as well. Subseries, in turn, are comprised of files, which subsequently consist of other files and/or items. The description of any single archival component inherits the description of its parents, grandparents, great-grandparents, and so on. Also inherited are layers or components of description that may reflect information shared by files within a given place in the hierarchy, but that do not necessarily correlate to an intellectual grouping to which the archival profession has assigned a specific name.

For example, in organizing a novelist's personal papers, you may identify several large groups or series of materials, one of which you name the Literary File and another Scrapbooks. Within the Literary File you may arrange the materials by type of writing such that all the files relating to the novelist's books are grouped together, as are the files relating to the writer's short stories and articles. Depending on the size and significance of the groupings, you may consider these three categories ("Articles," "Books," and "Short stories and other writings") as subseries and identify them as such in the finding aid. These subseries or types of writings may be further arranged by title, date, or another organizing principle, with additional subheadings applied as needed to complete the classification. The varying specificity of the classification often reflects the quantity of material in the category; the more material there is, the more likely the need for subcategories. Each category and subcategory of description is a component. A finding aid may depict this nesting as follows:(64)

	LITERARY FILE, 1943-1970, n.d.
		Articles, 1951-1966
		Books
			Raising Demons (1957)
				Reviews, 1956-1957, n.d.
				Royalty statements, 1956-1969
			The Road Through the Wall (1948), 1947-1970, n.d.
		Short stories and other writings
			"The Lottery"
				Dramatic adaptations
					Correspondence, 1949-1953, 1967-1970
					Scripts and screenplays, n.d.
				Royalty statements, 1950-1953, 1964-1970
			"Lover's Meeting," n.d.
	SCRAPBOOKS, 1933-1968
		College plays, 1933-1937
		"The Lottery," 1949-1952

In this much abbreviated and somewhat fictionalized example, the Literary File and the Scrapbooks are both series-level components in a collection of novelist Shirley Jackson's papers.(65) The Literary File component begins with the word "Literary" and ends after the file "Lover's Meeting." You might think that the component would end on the first line, after the span dates of the Literary File, but it does not; the description of the series-level component includes the description of all of its subcomponents, not simply the title and dates of the series. The component tags for the series would appear thus:

	<c>LITERARY FILE, 1943-1970, n.d.
		Articles, 1951-1966
		Books
			Raising Demons (1957)
				Reviews, 1956-1957, n.d.
				Royalty statements, 1956-1969
			The Road Through the Wall (1948), 1947-1970, n.d.
		Short stories and other writings
			"The Lottery"
				Dramatic adaptations
					Correspondence, 1949-1953, 1967-1970
					Scripts and screenplays, n.d.
				Royalty statements, 1950-1953, 1964-1970
			"Lover's Meeting," n.d.</c>
	<c>SCRAPBOOKS, 1933-1968
		College plays, 1933-1937
		"The Lottery," 1949-1952</c>

Both the Literary File and the Scrapbooks contain subcomponents, and these too would be tagged with beginning <c> and ending </c> component tags (see next example). In the Literary File, the archivist has identified three subcategories, two of which (books and short stories) have been further subdivided.

In determining how to encode a subcomponent, observe where its description begins and ends. In the case of the Jackson articles, the finding aid author decided that no further specification of the file's contents was needed other than a listing of the span dates. Consequently, the component begins before the word "Articles" and ends after the date range "1951-1966." With respect to the books and short stories, however, further levels of description were deemed appropriate to assist the researcher in locating material relating to a particular work by Jackson. Consequently, the books component begins just before "Books," but does not end until after the dates for The Road Through the Wall. Similarly, the short stories component begins immediately before "Short," and ends after the dates for "Lover's Meeting."

The process of identifying the beginning and end of each descriptive component would be followed throughout the descending hierarchy, as illustrated below. Note that the description of a lower-level component inherits the description of all its ancestors, such that a researcher reading the component title "Scripts and screenplays, n.d." will follow the hierarchy up and deduce that the file contains scripts and screenplays for dramatic adaptations of "The Lottery," which is a short story found in the literary file of the Shirley Jackson Papers.

	<c>LITERARY FILE, 1943-1970, n.d.
		<c>Articles, 1951-1966</c>
		<c>Books
			<c>Raising Demons (1957)
				<c>Reviews, 1956-1957, n.d.</c>
				<c>Royalty statements, 1956-1969</c></c>
			<c>The Road Through the Wall (1948), 1947-1970,
			n.d.</c></c>
		<c>Short stories and other writings
			<c>"The Lottery"
				<c>Dramatic adaptations
					<c>Correspondence, 1949-1953,
					1967-1970</c>
					<c>Scripts and screenplays,
					n.d.</c></c>
				<c>Royalty statements, 1950-1953,
				1964-1970</c></c>
			<c>"Lover's Meeting," n.d.
			</c></c></c>
	<c>SCRAPBOOKS, 1933-1968
		<c>College plays, 1933-1937</c>
		<c>"The Lottery," 1949-1952</c></c>

Note that regardless of their positions in a hierarchy, all components share the same generic element name and tag. Although each component identifies a hierarchically specific section of the described materials, no attempt was made in EAD to assign unique element names to various types or levels of component; all are simply tagged as <c>s. That does not mean that every <c> is of equal significance, or that an archivist may not wish to differentiate <c>s when displaying or searching a finding aid. Such differentiation is possible through the use of a LEVEL attribute, which, as explained in section 3.5.1., carries values of "collection," "fonds," "recordgrp," "series," "subgrp," "subseries," "file," "item," and "otherlevel." It is recommended that a LEVEL attribute be assigned for the highest <c>; the tags before "LITERARY FILE" and "SCRAPBOOKS" might therefore be expanded to <c level="series">. Elsewhere, the attribute may be used when the repository deems it useful for searching, display, navigation, or another purpose. Since many components are not likely to match any of the specific values available for the LEVEL attribute, little would be gained by identifying all of them as "otherlevel."

3.5.2.2. Unnumbered versus Numbered Components <c> and <c01>

Using the generic <c> to tag components, as shown in the above example, is perfectly valid, and most SGML authoring tools have no difficulty keeping track of the nested hierarchy when creating this encoded finding aid. Human comprehension may, however, be more problematic. When attempting to insert tags or proofread encoded documents prior to publication or dissemination, an archivist can easily become lost in a sea of <c> tags.

To address this authoring and editing problem, EAD offers an alternative encoding scheme of numbered components (<c01>, <c02>, <c03>, etc.), designed to assist an encoder in nesting up to 12 component levels accurately. Often the first series in a collection becomes the first <c01> in an EAD finding aid. The first <c02> is not the second series in the collection; rather, it is the next hierarchical level within that first <c01>.

It is crucial to realize that the numbers carry no intellectual significance and that their values are not absolute; in other words, a particular numbered <c> level invariably will correlate to a variety of intellectual levels, both within a single finding aid and across finding aids. For example, a <c02> in one part of a finding aid may be a file, while elsewhere, a <c02> may be a subseries. As with the unnumbered components, intellectual distinctions among components are made by using the LEVEL attribute. The example below illustrates the same section of the Jackson Papers finding aid seen earlier, but this time utilizing numbered components:

	<c01 level="series">LITERARY FILE, 1943-1970, n.d.
		<c02>Articles, 1951-1966</c02>
		<c02>Books
			<c03>Raising Demons (1957)
				<c04>Reviews, 1956-1957, n.d.</c04>
				<c04>Royalty statements, 1956-1969
				</c04></c03>
			<c03>The Road Through the Wall (1948), 1947-1970,
			n.d.</c03></c02>
		<c02>Short stories and other writings
			<c03>"The Lottery"
				<c04>Dramatic adaptations
					<c05>Correspondence, 1949-1953,
					1967-1970</c05>
					<c05>Scripts and screenplays, n.d.
					</c05></c04>
				<c04>Royalty statements, 1950-1953,
				1964-1970</c04></c03>
			<c03>"Lover's Meeting," n.d.</c03></c02></c01>
	<c01 level="series">SCRAPBOOKS, 1933-1968
		<c02>College plays, 1933-1937</c02>
		<c02>"The Lottery," 1949-1952</c02></c01>

Many EAD users feel that numbered <c>s are easier to use, but some prefer unnumbered <c>s, principally because less retagging of components is necessary if errors are spotted during editing or if an additional layer of description needs to be added at any level other than the lowest level. On the other hand, some EAD users have found that their search engines have been unable to process nested unnumbered <c>s, resulting in retrieval problems.

For this reason, these Guidelines recommend using numbered <c>s. Whichever approach is selected, you cannot switch between numbered and unnumbered <c>s within a single Description of Subordinate Components <dsc>, an element described in section 3.5.2.5.

3.5.2.3. Intellectual Content of Components

We have defined what components are and have explored the value of numbering them for proofreading and other purposes. We will next describe how to encode fully the intellectual content of individual components. Following the approach used in capturing information at the highest level of description, we will first focus on the use of the required <did> element and its subelements to ensure a sound basic description of each component <c>. Much of what was written about the <did> subelements in section 3.5.1.2 also applies to the component <c> descriptions and will not be repeated here. Tagged examples will illustrate how to utilize the full complement of elements previously described under <archdesc> (see section 3.5.1), while supplementary explanations will direct the reader's attention to specific attributes and uses not explored in the earlier <archdesc> discussion.
3.5.2.3.1. Basic Description of Each Component <did>
As noted earlier, information captured at the highest level of description is inherited by the subordinate components. Certain <did> subelements such as <repository> and <origination> therefore are rarely necessary at the component level. In contrast, other <did> subelements, such as <unittitle>, <unitdate>, <physdesc>, and <container>, are used frequently within components to encode more detailed description at a lower hierarchical level. These elements may be supplemented by others that are not <did> subelements, such as <scopecontent>, <arrangement>, and <organization>, or perhaps even <admininfo> and <bioghist>. A typical use of some of these elements is illustrated below in a series description taken from a Minnesota Historical Society finding aid for the records of that state's Game and Fish Commission, shown first without EAD encoding and then with sample tags:(66)

	Record of Prosecutions, 1916-1927.  3 volumes.
		Information provided in each entry: date of report, name and address of person
	arrested, locations where offense was committed, date of arrest, nature of offense,
	name of judge or justice, result of trial, amounts of fine and court costs, number
	of days served if jailed, name of warden, and occasional added remarks.  Types of
	offenses included hunting or fishing out of season or in unauthorized places,
	exceeding catch of bag limits, taking undersized fish, illegal fishing practices
	such as gill-netting or dynamiting, illegal hunting practices such as night-lighting,
	killing non-game birds, fishing or hunting without a license, and hunting-related
	offenses against persons such as fraud and assault.

 <c01 level="series">
	<did>
		<unittitle>Record of Prosecutions, <unitdate>1916-1927.
		</unitdate></unittitle>
		<physdesc>
			<extent>3 volumes.</extent>
		</physdesc>
	</did>
	<scopecontent>
		<p>Information provided in each entry: date of report, name and address of person
		arrested, locations where offense was committed, date of arrest, nature of offense,
		name of judge or justice, result of trial, amounts of fine and court costs, number
		of days served if jailed, name of warden, and occasional added remarks.  Types of
		offenses included hunting or fishing out of season or in unauthorized places,
		exceeding catch of bag limits, taking undersized fish, illegal fishing practices
		such as gill-netting or dynamiting, illegal hunting practices such as night-lighting,
		killing non-game birds, fishing or hunting without a license, and hunting-related
		offenses against persons such as fraud and assault.</p>
	</scopecontent>
</c01>

A series component from a manuscript collection, such as the Shirley Jackson Papers, would be encoded in a similar manner. The next example, however, illustrates not only the similarity in encoding descriptions of government records and personal papers, but also displays the unfolding hierarchy of EAD by showing how a description of a series component may be followed immediately by the description of its subcomponents.

	LITERARY FILE, 1943-1970, n.d.

	Correspondence, manuscript drafts, royalty statements, printed matter, notes,
	outlines, research material, screenplays, and miscellaneous items and
	enclosures relating to books and short stories by Jackson.  Organized
	alphabetically by type of material and arranged alphabetically by title or
	topic therein.  Publication dates of books are given in parentheses.

		Articles, 1951-1966
		Books
			Raising Demons (1957)
				Reviews, 1956-1957, n.d.
				Royalty statements, 1956-1969
			The Road Through the Wall (1948), 1947-1970, n.d.
		Short stories and other writings
			"The Lottery"
				Dramatic adaptations
					Correspondence, 1949-1953, 1967-1970
					Scripts and screenplays, n.d.
				Royalty statements, 1950-1953, 1964-1970
			"Lover's Meeting," n.d.


	<c01 level="series">
	<did>
		<unittitle>Literary File, <unitdate>1943-1970, n.d.
		</unitdate></unittitle>
	</did>

	<scopecontent>
		<p>Correspondence, manuscript drafts, royalty statements, printed matter,
		notes, outlines, research material, screenplays, and miscellaneous items and
		enclosures relating to books and short stories by Jackson.</p>
		<arrangement><p>Organized alphabetically by type of material
		and arranged alphabetically by title or topic therein.  Publication dates of
		books are given in parentheses.</p></arrangement>
	</scopecontent>

		<c02><did><unittitle>Articles, 1951-1966
		</unittitle></did></c02>
		<c02><did><unittitle>Books
		</unittitle></did>
			<c03><did><unittitle><title render="italic">
			Raising Demons </title>(1957)</unittitle></did>
				<c04><did><unittitle>Reviews, 1956-1957,
				n.d.</unittitle></did></c04>
				<c04><did><unittitle>Royalty statements,
				1956-1969</unittitle></did></c04></c03>
			<c03><did><unittitle><title render="italic">The
			Road Through the Wall </title>(1948), 1947-1970, n.d.
			</unittitle></did></c03></c02>
		<c02><did><unittitle>Short stories and other writings
		</unittitle></did>
			<c03><did><unittitle><title
			render="quoted">The Lottery</title></unittitle></did>
				<c04><did><unittitle>Dramatic adaptations
				</unittitle></did>
					<c05><did><unittitle>Correspondence, 1949-1953,
					1967-1970</unittitle></did></c05>
					<c05><did><unittitle>Scripts and screenplays, n.d.
					</unittitle></did></c05></c04>
				<c04><did><unittitle>Royalty statements,
				1950-1953, 1964-1970</unittitle> </did></c04></c03>
			<c03><did><unittitle><title render="quoted">
			Lover's Meeting, </title>n.d. </unittitle></did></c03>
		</c02>
	</c01>
3.5.2.3.1.1. Unit Title and Unit Date <unittitle> and <unitdate>
Note in the above examples that every component description includes the required <did> element, as well as a <unittitle>, which is optional but recommended within each <did>. It is also good descriptive practice to include the <unitdate> of each component. As mentioned in section 3.5.1.2.4, the <unitdate> may be placed either within or outside the <unittitle>. Many archivists in the United States (especially those who catalog using APPM) consider span dates to be part of the title, while descriptive practice in the United Kingdom dictates using <unitdate> outside <unittitle>. Each of the following examples is valid in EAD, but it is important that each repository select one method and use it consistently:

	<c04>
		<did>
			<unittitle><unitdate>1950-1961</unitdate></unittitle>
		</did>
	</c04>

	<c04>
		<did>
			<unitdate>1950-1961</unitdate>
		</did>
	</c04>

	<c04>
		<did>
			<unittitle>1950-1961</unittitle>
		</did>
	</c04>

Note that it is important not to use <head> as a substitute for <unittitle>. The <head> element is used to designate the title or caption for a section of text within the finding aid, such as "Contents List" or "Scope and Content Note." Never use <head> to identify the title of an archival unit or component. For example:

	CORRECT ENCODING:
		<c01 level="series"><did><unittitle>I.
		Correspondence</unittitle></did></c01>

The above encoding correctly identifies Correspondence as the title of the first series. The encoding shown below, which should not be used, simply creates a heading for the first component; it does not designate the intellectual content of the component:

	INCORRECT ENCODING:
		<c01 level="series"><head>I.
		Correspondence</head></c01>

Note that it is not necessary to use <head> to tag all text that you want to designate for display in a browser; a well-developed stylesheet should be able to pull in both <head>s and <unittitle>s for this purpose. See section 3.5.1.7.1 for information on the correct use of <head>.

As a general rule, do not use generic elements in any situation for which a specific element exists to designate the intellectual content of the described materials.

3.5.2.3.1.2. Physical Description <physdesc>
In the high-level <did> you will record the overall extent of the collection and may provide other general statements relating to the physical description of the whole (see section 3.5.1.2.5). In a component description, you are much more likely to take advantage of the four <physdesc> subelements:

Some of these elements may be further subdivided. For example, museums and archival repositories that hold three-dimensional objects or describe their materials at the item level will undoubtedly find the <physdesc> subelements extremely useful for encoding detailed item-level information.

The subelement <dimensions> may be used to record the size of an item in whatever unit of measurement the archivist selects. The UNIT attribute can specify the type of measurement, for example inches or centimeters, while the TYPE attribute can specify the kind of dimensions being measured, such as height or circumference.

	<dimensions>80 x 50 cm.</dimensions>

	<dimensions>
		<dimensions type="height" unit="inches">25</dimensions> x
		<dimensions type="width" unit="inches">38.5</dimensions>
	</dimensions>

The subelement <physfacet> encodes information about an aspect of the appearance of the described materials, such as their color, style, marks, substances, materials, or techniques and methods of creation. It is used especially to note aspects of appearance that identify or limit use of the materials. Again, the TYPE attribute specifies which aspect of the physical appearance is being designated. The <physfacet> element, like the controlled access elements described in section 3.5.3, also has a SOURCE attribute that identifies the thesaurus or other controlled vocabulary from which the term may be taken.

	<physfacet type="sealdesc">Rounded oval, a squirrel sitting on a tree,
	at the foot of which is a lion asleep.</physfacet>

	<physfacet type="medium">Paper, backed with linen.</physfacet>
<physfacet type="scale">4 chains to 1 inch.</physfacet>

See section 3.5.1.2.5 for examples illustrating use of <extent> and <genreform>. A very detailed <physdesc> at the item level might look like this:

	<physdesc>
		<physfacet type="medium">Parchment, with inserted paper leaves</physfacet>
		<extent>3 paper leaves; 1 parchment on paper leaf; 175 leaves,
		4 inserts, 2 schedules, parchment; 4 paper leaves</extent>
		<dimensions>230 x 163 mm.</dimensions>
		<physfacet type="binding">Modern, not later than 1824.</physfacet>
		<physfacet type="foliation">i-iv, 1-20, 20*, 21-37, 37*, 38-116,
		116*, 117-180 (Fifteenth century foliation: 1-84, on ff 125-158, 11-60).</physfacet>
	</physdesc>
	

3.5.2.3.1.3. Abstracts <abstract>
We noted in section 3.5.1.2.6 that an <abstract> within the high-level <did> is often extracted from longer descriptions found in <bioghist> and <scopecontent>. Within a component-level <did>, the abstract may describe particular characteristics of the individual component. This information usually has aspects of <arrangement>, <bioghist>, <physdesc>, and <scopecontent> that are not substantive enough to tag individually under those elements. Using <abstract> may be simpler, and is no less valid, than closing <did> and opening <scopecontent> and <p> outside <did>.

	<c02>
		<did>
			<container type="box">2</container>
			<unittitle>July 1925-June 1927</unittitle>
			<abstract>Includes an unexplained recapitulation of selected
			seizures for Oct. 1923 to 1925.  May have been people delinquent in
			paying fines.</abstract>
		</did>
	</c02>
	

3.5.2.3.1.4. ID Numbers and Physical Location Information <unitid> and <physloc>
In the high-level <did>, we noted that in addition to the recommended use of <repository>, <origination>, <unittitle>, <unitdate>, <physdesc>, and <abstract>, some repositories are also electing to include <unitid> and <physloc> with information about the collection as a whole (see section 3.5.1.2.7 and section 3.5.1.2.8). These same two elements also may be used at the component level if the information is different from that recorded in the high-level <did>. For example, a collection may be identified by a specific accession number or tracking number, and each component within that collection may be assigned a derivative of that parent number. Or the collection may consist of a group of lots, each of which is assigned a unique ID even if the collection as a whole does not carry any ID number.

	<c04>
		<did>
			<unittitle>Groups (includes Belafonte Singers), nos. 916-925 (F)</unittitle>
			<unitid>LOT 13074, nos. 767-987</unitid>
		</did>
	</c04>

In the high-level <did>, a building name, shelf location, or off-site storage address may be recorded in <physloc>. The same is true in <c><did><physloc>, which can be used to specify physical locations for individual components that are different from the rest of the collection, such as oversize items or films, sound recordings, or other types of materials that are likely to have special storage designations.

	<c06>
		<did>
			<unittitle><title render="italic">Women in Love
			</title>, </unittitle>
			<physdesc>page proofs </physdesc>
			<physloc>(Removed to galley file)</physloc>
		</did>
	</c06>

Most archival materials are housed inside file folders, boxes, and/or cabinets, and this storage locator information is often included in the finding aid. It is not, however, encoded using <physloc>. Another <did> subelement, <container>, is used to indicate the types of storage devices used and any sequential numbers assigned to those devices. Because this <container> information is of such a different character than most of the other information in the finding aid, section 3.5.2.4 is devoted to discussing its role and relationship to the intellectual markup.

3.5.2.3.2. Expanded Description of Components
As we noted several times in this chapter, all elements available to describe the collection as a whole are also available to describe the components. This means that, in addition to the <did> subelements, you may also use <admininfo>, <bioghist>, and <scopecontent> (discussed in sections 3.5.1.4 through 3.5.1.6). Also available are <arrangement> and <organization> (discussed in section 3.5.1.6 in connection with <scopecontent>, and illustrated at the component level in the last example under 3.5.2.3.1), as well as <add> (see section 3.5.4).

Any of these elements may be used whenever the description of the components differs from the parent collection or when the description given at the parent level needs additional detail. It may sometimes be appropriate to encode such data at the component level to which it most directly applies, such as the subgroup, series, file, or item. For example, information about restrictions may appear at the specific points in a container list where restricted items are described.

	<c01 level="series">
		<did>
			<unittitle>Search Files</unittitle>
		</did>
		<c02 level="file">
			<did>
				<unittitle>College of Engineering Dean Search
				<unitdate type="single">1986</unitdate>
				</unittitle>
				<physdesc>
				<extent>2 folders</extent>
				</physdesc>
			</did>
			<admininfo>
				<accessrestrict><p>Closed until <date type="restrict"
				normal="20100101">January 1, 2010</date></p></accessrestrict>
			</admininfo>
		</c02>
	</c01>

In such a situation, a high-level <accessrestrict> will explain the general nature of the access restriction and rolling opening dates. Specifically tagging the closure information will facilitate automated updating of finding aids as closure dates expire.

A <scopecontent> element at the <c> level might consist of a brief paragraph describing the topics and types of material reflected in an individual series. In other cases, an individual file or item may merit a longer description than is appropriate for the <abstract> element within <did>. In many cases, the decision to use <scopecontent> as opposed to <arrangement>, <organization>, or <abstract> will depend on the nature and length of text to be encoded. Aim for consistency in tag application within and across your finding aids, as such consistency will enhance both system design and researcher familiarity.

As mentioned in section 3.5.1.7.4, the <dao> and <daogrp> elements also are available at the <c> level. You might use these to embed or link to the contents of an entire file or selected items within a file, such as those that are frequently requested or that are fragile and cannot be handled in the original. Details on how to accomplish this are in section 7.3.6.

3.5.2.4. Physical Location and Container Information <container>

After describing the intellectual arrangement of a collection, you may add to this description other information for physically locating the materials within the repository. The materials typically reside in boxes and folders, or they may be reproduced on microfilm reels or stored in electronic form. This packaging or storage information, which sometimes may be a control or accession number rather than an actual container number, helps staff physically locate the particular files needed. Although in some cases this storage information may help researchers comprehend the size and extent of the materials relevant to their topics, it is principally of use to them only when submitting requests for materials.

Adding container or control number information expands the role of a finding aid by enabling it to serve not only as the access tool to the intellectual content of the materials, but also as the device for identifying their physical location. In traditional paper-based finding aids, these dual roles often are conveyed typographically through the use of columns. For example, an intellectual hierarchy may run down one side of a page, and a listing of container numbers, microfilm locations, or other control numbers may appear on the opposite side, as illustrated in the next example, in which container numbers have been added to the nested hierarchy previously shown. This container information is inherited in the same way that intellectual information is inherited. Even though each container number may appear only once in the finding aid, we assume the reader intuitively understands that the container number remains the same for all subsequent components until the next container number is listed. Within each container, folder numbers also are sometimes listed, creating a nested hierarchy of physical description parallel to the intellectual hierarchy.

These dual intellectual and physical hierarchies usually shift or break at different points. This is because a single container rarely holds all the materials described in a high-level component; it would be impractical to leave containers partially unfilled simply in order to isolate physically materials described as different intellectual components. By comparing the figure below with the preceding ones, you will see that the previously identified component descriptions span across containers, and that any single container may house materials that are intellectually part of different components. For example, the Literary File component begins in container 46 and ends in container 48, while the Books component starts in container 46 and ends at the beginning of container 47. Conversely, container 47 includes materials at various subcomponent levels within both Books and Short stories.

		LITERARY FILE, 1943-1970, n.d.
	46		Articles, 1951-1966
			Books
				Raising Demons (1957)
			Reviews, 1956-1957, n.d.
	47		Royalty statements, 1956-1969
				The Road Through the Wall (1948), 1947-1970, n.d.
			Short stories and other writings
				"The Lottery"
					Dramatic adaptations
						Correspondence, 1949-1953, 1967-1970
						Scripts and screenplays, n.d.
	48					Royalty statements, 1950-1953, 1964-1970
				"Lover's Meeting," n.d.

		SCRAPBOOKS, 1933-1968
	49		College plays, 1933-1937
	50		"The Lottery," 1949-1952

SGML does not accommodate simultaneous hierarchies effectively, and the EAD developers therefore had to choose which hierarchy the DTD would be optimized to handle. It seemed clear that intellectual arrangement is more important and more permanent than physical order, and that container information therefore should be secondary to the intellectual hierarchy rather than the reverse. In other words, the boxes and folders in which the materials are housed are not the components of a collection. The series, subseries, files, items, etc., are the components, and the container or packaging information is simply associated data that assists in retrieving those components. As a result, each new physical housing has no bearing on where a component begins or ends.

In EAD encoding, the container information "follows" the intellectual structure around; it is tagged as part of the description of the first component housed in that container. The example below illustrates this concept. It also illustrates the use of the TYPE attribute on <container>, the use of which is strongly recommended to clarify the nature of the storage device. Typical values include "box," "folder," and "reel." When an assigned number combines box and folder identification, the "box-folder" attribute value can be used. Please note that the example omits the <did> and <unittitle> tags required for valid markup in order to focus on the placement of the <container> tags.

	<c01 level="series">LITERARY FILE, 1943-1970, n.d.
		<c02><container type="box">46</container>
		Articles, 1951-1966</c02>
		<c02>Books
			<c03>Raising Demons (1957)
				<c04>Reviews, 1956-1957, n.d.</c04>
				<c04><container type="box">47</container>
				Royalty statements, 1956-1969</c04></c03>
			<c03>The Road Through the Wall (1948), 1947-1970, n.d.
			</c03></c02>
		<c02>Short stories and other writings
			<c03>"The Lottery"
				<c04>Dramatic adaptations
					<c05>Correspondence, 1949-1953, 1967-1970</c05>
					<c05>Scripts and screenplays, n.d.</c05></c04>
				<c04><container type="box">48</container>
				Royalty statements, 1950-1953, 1964-1970 </c04></c03>
			<c03>"Lover's Meeting," n.d. </c03></c02></c01>
	<c01>SCRAPBOOKS, 1933-1968
		<c02><container type="box">49</container>College plays,
		1933-1937</c02>
		<c02><container type="box">50</container>"The Lottery,"
		1949-1952</c02></c01>

The dispersal of components across containers is even more pronounced when you consider the scenario of the lowest-level component being housed in more than one container. The example below illustrates what might happen if there were five folders of reviews for Raising Demons, and all five could not be boxed in container 46. In this fictional case, the text appearing to the right of container 47 does not represent a new component, but merely an additional extent statement for the previous component ("Reviews, 1956-1957, n.d."). (Note: In converting existing finding aids, be careful not to assume that all typographical indentations represent the start of a new component. Sometimes, as in this example, indented text is simply a continuation of a previous component description.)

	Container Nos.	Contents

		LITERARY FILE, 1943-1970, n.d.

	46				Articles, 1951-1966
					Books
						Raising Demons (1957)
							Reviews, 1956-1957, n.d.
								(2 folders)
	47							(3 folders)
							Royalty statements, 1956-1969

The figure below shows how this section would be encoded to reflect that the <c04> component "Reviews" spans containers 46 and 47. Please note that the figure omits the necessary <did>, <unittitle>, and <physdesc> elements required for valid markup.

	<c01 level="series">LITERARY FILE, 1943-1970, n.d.
		<c02><container>46</container>
		Articles, 1951-1966</c02>
		<c02>Books
			<c03>Raising Demons (1957)
				<c04>Reviews, 1956-1957, n.d.
					<extent>(2 folders)</extent>
					<container>47</container><extent>
					(3 folders)</extent></c04>
				<c04>Royalty statements, 1956-1969</c04></c03>
				</c02></c01>

As the previous two encoded examples illustrate, EAD accommodates the traditional way in which container numbers have appeared in many finding aids-a single occurrence of the container number at the point in the finding aid where that container's contents begin. Archivists have believed that this minimalist approach has been sufficient and implicitly understood by readers. Unfortunately, as finding aids have moved into the electronic environment, a reader's ability to infer inherited container information has become more difficult. Whereas before a reader might have quickly skimmed a page of text or flipped back to a previous printed page, one must now scroll or jump through an electronic file, which may or may not display on the reader's computer as originally formatted by the creator.

In using EAD, it is important not to fixate on recreating the formatting of your existing paper guides. Some changes may be necessary to make finding aid information intelligible via electronic display to off-site users, who will not have ready access to a reference archivist. For example, some EAD users have concluded that to accommodate remote users, the container number should display for every component. Other archivists find this approach too labor intensive, dislike the repetitive display that results, and are unconvinced that the benefits are worth the increased file size caused by the additional tagging. As with any issue concerning the level of tagging employed, archivists need to consider not just the current display and navigation of online finding aids, but also whether they may want to manipulate and reuse information in an EAD document for various other purposes, both now and in the future.

One such immediate purpose, for example, would be to generate brief hit lists of materials across collections related to a specific topic or term. Although it would be helpful if this type of search returned to the reader a list of collection titles, imagine how pleased the same reader would be to receive a list that identifies not only the collection, but also the container location and specific component description within that collection. The ease with which a computer system can deliver such results will depend to some degree on the amount of EAD encoding explicitly provided, since computers are not able to interpret the many implicit connections between data that humans routinely make. While a human may deduce that a box number applies to all subsequent items until a new box number is given, a computer will have a far more difficult time making such an inference. Therefore, for the sake of machine-processing, you may wish to consider encoding the container and/or folder information for each component. As an alternative, you may wish to explore using the PARENT attribute, which is available on the <container> and <physloc> elements (see section 7.2.5.)

As with many aspects of EAD implementation, you should weigh the anticipated value of encoding container information for every component against the cost of doing so. Also consider the system's ability to suppress information not desired for display or retrieval. For example, you may want the container information to display in the brief "hit list" format, but not on every line of the electronic finding aid. One could probably write a stylesheet that would cause only the first instance of a physical location and container to display, but the ability to support such conditional scripting will likely vary from browser to browser. A more appealing (but more difficult) approach would be for the system to generate a dynamic or running header at the top of the screen that would display container location and higher-level component data that change as you move through the finding aid. Since the ability to generate such displays may be dependent upon the level of encoding, archivists should consider encoding data more extensively than current software supports or than existing needs require.

3.5.2.5. Formatting Component-Level Descriptions: The Description of Subordinate Components <dsc>

The previous sections have shown how an EAD finding aid can capture the unfolding hierarchical structure of an archival collection, in which the description of each component inherits the description of the higher-level components that have preceded it. Think of it, for a moment, as a steady but changing stream of water, in that the information flows downward from the <archdesc> to the various <c>s, with fluctuations in depth occurring as the number of nested components rises and falls throughout the hierarchy. As the information courses its way downward from the whole to the parts, it must first, however, pass through another element, the Description of Subordinate Components <dsc>. The <dsc> element is a formatting construct, not unlike an aqueduct, that redirects the information flow along several possible routes in an attempt to accommodate the anticipated navigation of information along the way.

The <dsc> element must be opened before any <c> element is accessible. The <dsc> is a required wrapper that surrounds the subordinate component <c> elements within <archdesc> and helps alert the computer to the special formatting and structural features that may appear as the information stream changes course. As discussed in previous sections, the description of the whole generally consists of short phrases of text in the high-level <did> subelements, followed by longer prose statements encoded as <admininfo>, <bioghist>, <scopecontent>, <organization>, and <arrangement>. For most of the latter elements, the content model is a <head>, which serves to help users interpret the data, followed by text composed of <p>s or various types of <list>. Although <admininfo>, <bioghist>, <scopecontent>, and other narrative-based elements are also available at the <c> level, traditional finding aid practice is such that most component-level descriptions consist of relatively brief <did>-like information arranged in indented columns or outlines.

The need to accommodate such legacy formats led the EAD developers to create the <dsc> in order to meet several needs:

A TYPE attribute is required with <dsc> to identify the direction or course the information stream will follow, and four possible values are available:

The most natural flow (and the one that these Guidelines recommend) is to set the TYPE attribute to "combined." This signals an information sequence similar to figure 3.5.2.5a, in which a description of a series or subseries is immediately followed by the descriptions of its component parts.

LITERARY FILE, 1943-1970, n.d.

Correspondence, manuscript drafts, royalty statements, printed matter, notes,
outlines, research material, screenplays, and miscellaneous items and enclosures
relating to books and short stories by Jackson.  Organized alphabetically by type
of material and arranged alphabetically by title or topic therein.  Publication
dates of books are given in parentheses.

	Articles, 1951-1966
	Books
		Raising Demons (1957)
			Reviews, 1956-1957, n.d.
			Royalty statements, 1956-1969
		The Road Through the Wall (1948), 1947-1970, n.d.
	Short stories and other writings
		"The Lottery"
			Dramatic adaptations
				Correspondence, 1949-1953, 1967-1970
				Scripts and screenplays, n.d.
			Royalty statements, 1950-1953, 1964-1970
		"Lover's Meeting," n.d.

	Figure 3.5.2.5a.  Combined Model <dsc type = "combined">
		Formatted Description of Series.


<dsc type="combined">
	<c01 level="series">
		<did>
			<unittitle>Literary File, <unitdate>1943-1970,
			n.d.</unitdate></unittitle>
		</did>

		<scopecontent>
			<p>Correspondence, manuscript drafts, royalty statements, printed
			matter, notes, outlines, research material, screenplays, and miscellaneous items
			and enclosures relating to books and short stories by Jackson.</p>
			<arrangement><p>Organized alphabetically by type of material
			and arranged alphabetically by title or topic therein.  Publication dates of
			books are given in parentheses.</p></arrangement>
		</scopecontent>

		<c02><did><unittitle>Articles, 1951-1966
		</unittitle></did></c02>
		<c02><did><unittitle>Books</unittitle></did>
			<c03><did><unittitle><title
			render="italic">Raising Demons </title>(1957)
			</unittitle></did>
				<c04><did><unittitle>Reviews, 1956-1957,
				n.d.</unittitle></did></c04>
				<c04><did><unittitle>Royalty statements,
				1956-1969</unittitle></did></c04></c03>
			<c03><did><unittitle><title render="italic">
			The Road Through the Wall </title>(1948), 1947-1970,
			n.d. </unittitle></did></c03></c02>
		<c02><did><unittitle>Short stories and other
		writings </unittitle></did>
			<c03><did><unittitle><title render="quoted">The Lottery
			</title></unittitle></did>
				<c04><did><unittitle>Dramatic adaptations
				</unittitle></did>
					<c05><did><unittitle>Correspondence, 1949-1953,
					1967-1970</unittitle></did></c05>
					<c05><did><unittitle>Scripts and screenplays,
					n.d.</unittitle></did></c05></c04>
				<c04><did><unittitle>Royalty statements, 1950-1953,
				1964-1970</unittitle></did> </c04></c03>
			<c03><did><unittitle><title render="quoted">Lover's
			Meeting, </title>n.d. </unittitle></did></c03>
		</c02>
	</c01>
</dec>

	Figure 3.5.2.5b.  Combined Model <dsc type = "combined">
		Encoded Description of Series.


The analytic overview model (type="analyticover") is for encoding a high-level component overview, such as when each of a collection's series (and possibly subseries) is individually described at its highest level, one right after the other, for collective display (see figure 3.5.2.5c and figure 3.5.2.5d). In other words, a <c01> is opened, the first series is briefly described, and then the <c01> is closed without its subcomponents yet having been described. Then the second <c01> is opened, the second series is briefly described, and this <c01> is closed, to be followed in the same manner by all subsequent <c01> components. As in the "combined" model, the descriptive overview of each series is encoded using the component <c> or <c01> tags, with the LEVEL attribute set to "series."

The in-depth model (type="in-depth") is designed to encode a contents list separate from any narrative series descriptions. Each series, subseries, file, or item represented in the contents list is tagged as a recursive, nested component, possibly with an optional LEVEL attribute set to identify its hierarchical order within the collection or record group. (See figure 3.5.2.5e and figure 3.5.2.5f)

The last <dsc> value, othertype (type="othertype") is for use with models that do not follow any of the three specifically defined formats, to which you may assign any appropriate name.

More than one type of <dsc> may be used within a single EAD finding aid. Encoding a <dsc type="analyticover"> followed by a <dsc type="in-depth"> is a common approach found in many legacy finding aids. This "two-<dsc> approach" has both advantages and disadvantages. It accommodates a legacy data structure without requiring the shifting or rekeying of text. In addition, it preserves the functionality that the legacy structure provided by assembling in one spot all the first-level component descriptions so that they may be quickly scanned by a researcher. It would be less convenient to flip through a long paper guide, or to scroll and jump through an electronic finding aid, to locate all the first-level summaries encoded under the combined <dsc>.

On the other hand, the two-<dsc> approach has drawbacks. Because you are encoding the same first-level components (<c01>) more than once (first in the analytic overview and then in the in-depth presentation), you are creating potential confusion for a computer trying to process identical information. Depending on the sophistication of a system's searching and processing capabilities, the two-<dsc> approach may hamper the ability to show a relationship between the description of the <c01> and the description of its parts. The next generation of Web browsers may offer a solution by accommodating scripts that would locate and display or print together all the series-level descriptions in a combined <dsc>. In such a scenario, the functionality of the analytic overview could be preserved without having to split a collection's natural hierarchical description between two <dsc>s.

In the meantime, archivists using the two-<dsc> approach may want to provide hypertext links from the description of the series in the analytic overview to the description of its components in the in-depth <dsc> in lieu of encoding the same first-level components twice.

 Container Nos.		Series

1		Diary and Diary Notes, 1932-1934, n.d.
			A high-school diary and an undated, single-page diary fragment
			kept by Jackson.  Arranged chronologically.

2		Family Papers, 1938-1965, n.d.
			Letters received, notes, and cards.  Arranged alphabetically
			by family member and chronologically therein.

3-12	Correspondence, 1936-1970, n.d.
			Letters received and occasional copies of letters sent,
			telegrams, postcards, and miscellaneous enclosures.  Arranged
			alphabetically by correspondent and chronologically therein.

13-19	Literary File, 1943-1970, n.d.
			Correspondence, manuscript drafts, royalty statements, printed
			matter, notes, outlines, research material, screenplays, and
			miscellaneous items and enclosures relating to books and short
			stories by Jackson.  Arranged alphabetically by type of material
			and alphabetically by title or topic therein.  Publication dates
			of books are given in parentheses.

	Figure 3.5.2.5c.  Analytic Overview Model <dsc type = "analyticover">
		Formatted Description of Series.


<dsc type="analyticover">
	<head>Description of Series</head>
	<thead><row><entry>Container Nos.</entry>
	<entry>Series </entry></row></thead>

	<c01 level="series">
		<did>
			<container>1</container>
				<unittitle>Diary and Diary Notes, <unitdate>
				1932-1934, n.d. </unitdate></unittitle>
		</did>
		<scopecontent><p>A high-school diary and an undated,
		single-page diary fragment kept by Jackson.</p>
		<arrangement><p>Arranged chronologically.</p>
		</arrangement>
		</scopecontent>
	</c01>
	<c01 level="series">
		<did>
			<container>2</container>
			<unittitle>Family Papers, <unitdate>
			1938-1965, n.d. </unitdate></unittitle>
		</did>
		<scopecontent><p>Letters received, notes, and cards. </p>
		<arrangement><p>Arranged alphabetically by family
		member and chronologically therein.</p></arrangement>
		</scopecontent>
	</c01>
	<c01 level="series">
		<did>
			<container>3-12</container>
			<unittitle>Correspondence, <unitdate> 1936-1970, n.d.
			</unitdate></unittitle>
		</did>
		<scopecontent><p>Letters received and occasional copies
		of letters sent, telegrams, postcards, and miscellaneous enclosures. </p>
		<arrangement><p>Arranged alphabetically by correspondent
		and chronologically therein.</p></arrangement>
		</scopecontent>
	</c01>
	<c01 level="series">
		<did>
		<container>13-19</container>
		<unittitle>Literary File, <unitdate>1943-1970, n.d.
		</unitdate></unittitle>
		</did>
		<scopecontent><p>Correspondence, manuscript drafts, royalty
		statements, printed matter, notes, outlines, research material, screenplays,
		and miscellaneous items and enclosures relating to books and short stories
		by Jackson. </p>
		<arrangement><p>Arranged alphabetically by type of material
		and alphabetically by title or topic therein.</p></arrangement>
		<p>Publication dates of books are given in parentheses.</p>
		</scopecontent>
	</c01>
</dsc>

	Figure 3.5.2.5d.  Analytic Overview Model <dsc type = "analyticover">
		Encoded Description of Series.


LITERARY FILE, 1943-1970, n.d.
Container Nos.	Contents

46				Articles, 1951-1966
				Books
					Raising Demons (1957)
						Reviews, 1956-1957, n.d.
						Royalty statements, 1956-1969

47					The Road Through the Wall (1948), 1947-1970, n.d.
				Short stories and other writings
					"The Lottery"
						Dramatic adaptations
							Correspondence, 1949-1953, 1967-1970
							Scripts and screenplays, n.d.

48						Royalty statements, 1950-53, 1964-70
					"Lover's Meeting," n.d.

	Figure 3.5.2.5e.  In-Depth Model <dsc type = "in-depth">
		Formatted Container List.


<dsc type="in-depth">
	<head>Container List</head>
	<thead><row valign="top"><entry colname="1">
	Container Nos.</entry><entry colname="2">Contents
	</entry></row></thead>

	<c01 level="series">
		<did>
			<unittitle>LITERARY FILE, <unitdate
			type="inclusive">1943-1970, n.d.</unitdate>
			</unittitle>
		</did>
		<c02><did><container>46</container>
		<unittitle>Articles, 1951-1966</unittitle></did></c02>
		<c02><did><unittitle>Books</unittitle></did>
			<c03><did><unittitle><title render="italic">
			Raising Demons</title> (1957)</unittitle></did>
				<c04><did><unittitle>Reviews, 1956-1957, n.d.
				</unittitle></did></c04>
				<c04><did><unittitle>Royalty statements, 1956-1969
				</unittitle></did></c04>
			</c03>
			<c03><did><container>47</container>
			<unittitle><title render="italic">The Road Through
			the Wall</title> (1948), 1947-1970, n.d. </unittitle>
			</did></c03></c02>
		<c02><did><unittitle>Short stories and other
		writings </unittitle></did>
			<c03><did><unittitle><title render="quoted">
			The Lottery</title></unittitle></did>
				<c04><did><unittitle>Dramatic adaptations
				</unittitle></did>
					<c05><did><unittitle>Correspondence,
					1949-1953, 1967-1970</unittitle></did></c05>
					<c05><did><unittitle>Scripts and screenplays,
					n.d.</unittitle> </did></c05></c04>
				<c04><did><container>48</container>
				<unittitle>Royalty statements, 1950-1953, 1964-1970
				</unittitle></did></c04>
			</c03>
			<c03><did><unittitle><title render="quoted">
			Lover's Meeting, </title> n.d.</unittitle></did>
			</c03>
		</c02>
	</c01>
	[...]
</dsc>

	Figure 3.5.2.5f.  In-Depth Model <dsc type = "in-depth">
		Encoded Container List.

Continue

Back

Return to Menu


Footnotes

  1. The term "file" is used here to mean an intellectual unit of archival materials, not the contents of a physical housing such as a folder.

  2. All examples in section 3.5.2 and its subsections are adapted from: Janice E. Ruth, "Encoded Archival Description: A Structural Overview," American Archivist 60 (summer 1997): 310-29.

  3. This collection is held by the Library of Congress.

  4. Examples throughout this section on components do not necessarily include all possible or even all required elements. The examples are intended to illustrate the use of certain specific elements, which full tagging might obscure. Note also that date information is not specifically encoded below the series (<c01>) level. Because such a variety of date formats are used in finding aids (see also section 3.5.1.2.4), the encoding of date information at the file or item level is of questionable utility for retrieval purposes. See appendix E for fully tagged examples.


Table of Contents
Home Page Preface Acknowledgments How to Use
This Manual
Setting EAD
in Context
Administrative
Considerations
Creating Finding
Aids in EAD
Authoring EAD
Documents
Publishing EAD
Documents
SGML and XML
Concepts
EAD Linking
Elements
Appendices


Go to:


Copyright Society of American Archivists, 1999.
All Rights Reserved.


[VIEW OF LC DOME] The Library of Congress

Library of Congress Help Desk (11/01/00)