Abstract

Information systems are based on the two basic principles of structured information and well-defined ways of handling this information. While the Web as an information system has a number of well-designed ways to handle information, the foundation of Web-based applications, how information should be structured, are less clear. When considering information structures, the most important aspects are the question of classification systems and how to apply these to existing information, and the question of how classified information can be composed into structures which can be regarded as useful and self-contained representations for relevant information.

Computers and Information

• Information systems need formalized structures
  • if computers should act on information, there need to be well-defined structures
• Structures appear on various levels
  • information atoms need to be conform to a structure
  • atoms need to be composed into bigger structures
  • structured items can have a variety of relationships
• Alternative approaches are less rigid in their need for structure
  • neural networks are trained rather than programmed and find structure
  • data mining finds structures by associating large data sets
  • signal processing finds patterns in signals (image, audio, video)

Names

• A name is a label for some thing or some category that is used to distinguish one from another
• A thing or category can often have multiple names; these are synonyms or aliases
• If the name is unique in some context it can be called an identifier
  • this means identifiers are always names, but not vice versa
• Different things can sometimes have the same names, these are homonyms or polysemes
  • homonyms simply have the same syntax or pronunciation (homographs must have the same syntax)
  • polysemes share a deeper commonality rooted in etymology

Controlled Vocabularies

• A controlled vocabulary is a standardized set of terms
  • assigned by organizers of information
  • to help searchers find information (with high recall and precision) and to identify the entity found
• It is a content standard for use in (or as) metadata elements
• Its goal is to impose some order to facilitate agreement between the concepts within the site and the vocabulary of the person [natural language] using it
• It can be thought of as a fixed or closed dictionary in which everything must be defined using the same set of terms

Vocabulary Stability

• Vocabulary control for places
  • variant forms: St. Petersburg, Санкт Пербургскйй, Saint-Pétersbourg
  • multiple names: Cluj, in Romania/Roumania/Rumania, is also called Klausenburg and Kolozsvar
  • name changes: Bombay → Mumbai
  • homographs: Vienna, VA, and Vienna, Austria; 50 Springfields
  • anachronisms: no Germany before 1870
  • vague: e.g. Midwest, Silicon Valley
  • unstable boundaries: 19th century Poland; Balkans; USSR
• Maintaining controlled vocabularies is not easy
  • a lot of decisions must be made which could be made one way or the other
  • users of the vocabulary must agree to accept these decisions
• Controlled vocabularies are artificial languages
  • created to to facilitate agreement in some context and for some purposes

Classification

• A classification is a system of categories
  • categories are equivalence classes, ie.e. sets entities that we treat the same
  • classification (verb) is the process of assigning entities to the categories in a classification system
• It is a reference model to individual domains and the relationships among them to help people understand the concepts and relationships
• It improves learning and communication
• It supports information retrieval and interfaces for information systems
• Classification is lawful, systematic, and arbitrary
  • lawful because it follows a set of principles that govern the structure of the categories and their relationships
  • systematic because it requires that those principles be followed
  • arbitrary because the criteria used to establish the categories reflect a perspective on the domain that intentionally excludes all other perspectives

Library of Congress Classification

A -- GENERAL WORKS
B -- PHILOSOPHY. PSYCHOLOGY. RELIGION
C -- AUXILIARY SCIENCES OF HISTORY
D -- HISTORY (GENERAL) AND HISTORY OF EUROPE
E -- HISTORY: AMERICA
F -- HISTORY: AMERICA
G -- GEOGRAPHY. ANTHROPOLOGY. RECREATION
H -- SOCIAL SCIENCES
J -- POLITICAL SCIENCE
K -- LAW
L -- EDUCATION
M -- MUSIC AND BOOKS ON MUSIC
N -- FINE ARTS
P -- LANGUAGE AND LITERATURE
Q -- SCIENCE
R -- MEDICINE
S -- AGRICULTURE
T -- TECHNOLOGY
U -- MILITARY SCIENCE
V -- NAVAL SCIENCE
Z -- BIBLIOGRAPHY. LIBRARY SCIENCE. INFORMATION RESOURCES (GENERAL)

Biased Classification

200 Religion
        210 Natural theology
        220 Bible
        230 Christian theology
        240 Christian moral & devotional theology
        250 Christian orders & local church
        260 Christian social theology
        270 Christian church history
        280 Christian sects & denominations
        290 Other religions

• Every classification is biased
  • every classification system takes a point of view
  • distinction between good or standard and nonstandard ways of understanding things

Hierarchical Literature Classification

Faceted Literature Classification

Faceted Classifications

• Relationships between facets enable a small controlled vocabulary to generate
  • many structured descriptions
  • that are complex, but formally structured
  • that enable us to describe things we do not have words for
• Faceted classification enables very effective user interfaces for browsing and searching in structured collections

Criteria for Facets

• Orthogonality: facets are independent dimensions
• Semantic Balance: top level facets are the most important semantic dimensions of the domain; values within facets are at equal semantic level
• Coverage: all current instances can be classified
• Scalability: future instances can be classified
• Objectivity: instances can be objectively classified
• Not idiosyncratic: facet semantics should be mainstream or normative

Ontologies

Ontologies are a formalization of a conceptualization.

• Ontologies define relationships between terms
  • classification systems are a very weak form of ontology
• Ontologies may only define lexical relationships between terms
  • synonyms and other forms of lexical relations between classification terms
• Ontologies often define semantic relationships between terms
  • depending on the formalism, very complex relations can be expressed
  • a child always has two parents, a father and a mother
  • a child's child is called a person's grandchild
  • two children sharing exactly one parent are called step-siblings

Composition

• Information often needs to be grouped into compound entities
  • the atomic entities are free form or classified
  • there is some relationship that has to be captured
• Documents are composed out of individual parts
  • some documents have only free form content
  • some documents have only classified content
  • most documents are a mixture of these two extremes
• Content management is the act of managing documents
  • more classified content means more publishing and management capabilities

What is a Document?

Document Definition

• A document is a purposeful and self-contained collection of information
• This definition focuses on the information content, not on the physical container or medium, format, or technology in which the information is conveyed
• This broader definition includes information extracted from databases or applications, spreadsheets, printed or Web forms, as well as traditional printed documents

Document Types

• Any definition of document allows for a notion of different types of documents
• This idea can be very intuitive and very informal, or we can be more precise and define a model of a document type as the rules or constraints that distinguish one type from another
• This expression of the model is conceptual and is independent of the syntax and technology in which document instances are ultimately implemented
• But most of the time the model is ultimately implemented as a physical view in some specific syntax like XML

Models of Document Types

• A model of a document type captures the distinctions between documents that make a difference
• Similar types of content occur in many document models and there is often overlap in information and structural patterns
• Models of document types can be very specific (purchase order for industrial chemicals when buyer and seller are in different countries) or very abstract (fill-in-the-blank legal form for contract)
• The constraints on structure, data, and behavior of the information in the conceptual and physical models of a document type can collectively be called the Business rules

Business Rules

• Business Rule is an up and coming buzzword in information systems that is a lot broader than its name might suggest
• A Business Rule is a formal or refined statement of an requirement that is expressed in a technology-independent fashion

Business Rule Taxonomy

• Rules that apply to conceptual models
  • semantic (properties, generalizations, dependencies)
  • structural (co-occurrences, aggregations)
  • usage (policies, roles, responsibilities)
• Rules that (can also) apply to physical models
  • syntactic (encoding format, schemas, character sets)
  • processing (procedural or behavioral rules, protocol issues)
  • presentational (layout information, accessibility)
• Rules that apply to instances
  • content (constraints)
• Concrete rules often mix these types

Business Rule Patterns

• Document types differ in the mixture of business rules from these seven categories
• Put another way, the mixture of rules determines the kind of document type
  • Transactional documents have many semantic, usage, processing, and content rules
  • Narrative documents tend to have more structural and presentational rules
• The mixture of these business rules varies in systematic ways across the spectrum of document types

Document Type Spectrum

Document Type Examples

• Narrative Document Types (or Publications)
  • examples: Novels, user guides, technical manuals, ...
• Transactional Document Types
  • examples: Orders, invoices, payment instructions, ...
• In between are Hybrids
  • examples: Dictionaries, catalogs, ...

Documents, Rules, and Schemas

• Rules that define a document type in a form that is computable or processable enable:
  • the creation of valid document instances in editors or when information is exported from a database or other application
  • a model for application programming in the development of Web forms or other GUIs or a template for objects in other programming environments
  • communication of the model to others who need to create or receive document instances
• XML schemas attempt to encode the conceptual model in terms of the syntactic constructs of elements and attributes
  • what elements and attributes are allowed (vocabulary)
  • where the are allowed (grammar) – sequence, choice, occurrence and co-occurrence
  • What values they can take (datatypes) – string, integer, decimal, etc.

Schema Languages

• Schema languages are the formalisms in which schemas are described
• They differ substantially in how completely they can express the business rules that underlie a model of a document type
• No schema language is perfect; there is always some compromise between:
  • Expressiveness – the range of models that can be described
  • Functionality – the set of features used to define a model
  • Usability – the ease with which a model can be defined
  • Reusability – how readily a model or parts of models can be included in another one
  • ...and a range of other ...ilities
• XML has a variety of schema languages associated with it
  • Document Type Definition (DTD) are a part of XML itself and thus an important part of XML
  • XML Schema is a separate standard defining a more powerful and complex language

Documents Represent Information

• Data needs to be classified to become information
• Information can be exchanged by using documents
• Mapping document information to system information creates knowledge