The Internet Encyclopedia (Volume 3)

P1: 61

WL040C-197-Quin WL040/Bidgoli-Vol III-Ch-56 June 23, 2003 16:38 Char Count= 0

CONTENTREPRESENTATION ANDORGANIZATION 691

Content representation and organization refers to the

process in which data and document models are created

and relationships are mapped out as the architecture for

templates and interfaces. Subsequently, such models are

converted into logic designs and implemented into the sys-

tem. Creation of representation and organization schemes

and vocabularies requires methodologies from computer

science, linguistics, library and information science, and

other allied disciplines. There are two kinds of tools in

content representation and organization: (1) metadata

schemes that contain data elements describing various

types of data sets and documents and (2) controlled vo-

cabularies that are used to assign subject categories and

indexing terms to documents.

Metadata Schemes

A metadata scheme defines a set of properties about data

or documents. In the digital library community, there

have been several proposals: the Dublin Core metadata

sets (Weibel, Kunze, Logoze, & Wolf, 1998), the War-

wick Framework (Lagoze, 1996), and the IEEE LOM/IMS

metadata set for learning objects (IEEE/LTSC, 2002; IMS,

2002). The main purpose of metadata schemes is to pro-

vide a consistent way to record data about data sets

or documents and to encode such data in a computer-

understandable manner. The most common metadata

elements include authors, titles, dates of creation/

publication, owner/publisher, description, and category/

subject terms. Some metadata schemes provide coarse-

grained metadata elements to allow quick generation or

creation of a metadata repository. Dublin Core is an exam-

ple of this kind. With a structure containing three pack-

ages of content, intellectual property, and instantiation,

it includes 15 elements and their subelements (Table 1).

This set enumerates the very general yet essential data el-

ements necessary for identifying and locating documents.

Metadata schemes may also include finer-grained data

elements to describe documents in greater detail. For ex-

ample, Dublin Core’s educational extension, DC-ED, em-

ploys a set of elements designed specifically for describing

the grade level, audience, pedagogy, and other details for

a learning resource on the Web. In this sense, metadata

schemes are the data models for collecting information

about documents on the Web. Such information is used

to search, browse, and locate documents on a topic over

the Web. Another use of metadata schemes is to design

them as templates to capture content directly from the

Table 1Elements in Dublin Core

Content Intellectual property Instantiation

Title Creator Date

Subject Publisher Type

Description Contributor Format

Source Rights Identifier

Language

Relation

Coverage

Web. An example is Web-based forms, which have been

widely used to capture customer data, survey data, sales

orders, and best practices, to name a few.

Sometimes no standard metadata scheme is suitable

for representing the types of data and documents being

created. In such circumstances, a customized metadata

scheme(s) is needed. The customized scheme may be cre-

ated based on a standard such as Dublin Core, with a set

of extended elements. It may also be created from scratch

without using any existing standard. However, for inter-

operability, a metadata standard should be adopted when-

ever possible. Since metadata schemes are the data mod-

els for collecting information about data and documents,

they often need to be embedded in content authoring tools

so that content authors can create metadata while they

are creating documents. In either case, i.e., creating an ex-

tended metadata scheme or a scheme for capturing data in

a distributed environment, careful design is necessary to

guarantee a sound representation of the data to allow easy

retrieval and browsing of the content. Designing metadata

schemes usually involves a data modeling process.

Modeling is a process of synthetic analysis and abstrac-

tion, in which system developers “construct an abstract

description of a system in order to explain or predict cer-

tain system properties or phenomena” (Schreiber, Akker-

mans, Anjewierden, Hoog, Shadbolt, & Wielinga, 2000, p.

128). The result of this process is a model in which classes

of objects in the real world are specified as having a trans-

parent and one-to-one correspondence to an object in the

model. Figure 3 is an example of the data model for an ex-

perience factory in the domain of workforce development.

The model in Figure 3 provides a high-level view of

main classes and the relations among the classes. In the

domain of workforce development, participating classes

include the government who sponsors workforce pro-

grams such as “Welfare-to-work” and “Study-to-career,”

workforce organizations that initiate projects to execute

programs, people in organizations who prepare docu-

ments describing and disseminating project information

and results, and knowledge captured for sharing promis-

ing practices and lessons learned. A data model such as

this serves as a communication tool for users of the Web

content. It shows what kinds of content there will be on

the Web site and how each class is related to one another.

Before a data model reaches its acceptable version, discus-

sions are often held with constituents to solicit feedback.

As with any system development process, this discussion–

revision process is iterative.

Cases

that worked

Lessons

learned

Workforce

Document Projects programs

People Organization Government

describes refers-to

example-of

includes is-related-to

writes initiates is-part-of sponsors

Figure 3: A sample conceptual model for the workforce ser-

vices domain.