Abstract

Web-based publishing is based on the assumption that there is some data source providing the data which is then published in a Web-compliant way. It is the task of a Content Management System (CMS) to provide an platform for creating, managing, repurposing, and publishing content in a flexible way. CMS are often based on XML, because it provides a good foundation for flexible reuse, but this still leaves open the question of the schema and the available publication pipelines.

What is Content Management?

• What semantics are necessary to be able to reuse content?
• What structures are necessary to be able to reuse content?
• What publishing pipelines need to be available?
• What is the content lifecycle (or content management workflow)?
• What about scalability (content size, throughput)?
• What about the legal issues of storing and providing content?

What is a CMS?

• Requirements in publishing environments are very different
  • highly structured document models and document workflows are harder to use
  • less structured models and workflows lead to lower-quality content
  • CMS applications and systems are operating in a continuum
• Highly structured models and workflows for high-quality data
  • content creation is hard and may suffer from spurious data entry
  • highly complex and finely grained structures allow very rich content structures
  • this side of the spectrum is leaning towards the read-only extreme
• Less structured models and workflows for lower barrier to entry
  • content creation is easy but only little structure and metadata is captured
  • wikis have become en vogue, even as enterprise wikis
  • this side of the spectrum is leaning towards the write-only extreme

ETH University CMS

• Web sites at ETH evolved with no central support or guidance
  • a lot of effort and expertise duplicated (a couple of hundred Web sites)
  • no corporate design, no quality of service, a lot of redundant data
• ETH World was a program for advanced IT development
  • funded by the university and granted to various IT-related projects
  • Web-related issues were considered strategic communications
  • online newspaper, corporate design, centralized CMS, consolidated services
  • April 2007 statistics: 339 active Web sites, 125 under development
• Silva is now part of the regular services of the ETH IT department
  • open source project with outsourced development of required functionality
  • Silva is now being used by a number of organizations

ETH Home Page

CSG Home Page

Suboptimal Web Design

Migrating Sites

• Nobody is forced to migrate to the new design and CMS
  • migration naturally occurs if the CMS is better than homegrown sites
  • additional features in the CMS should make migration attractive
• Web site management for a long time had no strategic home
  • Web sites were built locally with a lot of effort and emotional attachment
  • many Web sites built some form of content management (calendar, phone book, publications, …)
  • migrating a Web site sometimes affects the whole IT of a department
• The more CMS services ETH provides, the more users will migrate
  • Silva can include content from the course catalog (requires advanced caching)
  • there is a calendar service (but only for university-wide dates)
  • there is no publication data (a lot of proprietary solutions have been developed)

Identities and Roles

• Users are the most basic concept for identification and authentication
  • groups allow to classify users according to application needs
  • groups are not necessarily hierarchical
  • users may belong to more than one group
• Roles are a concept similar to groups
  • groups often represent organizational structures
  • roles often represent duties and permissions for certain application areas
• Delegation is crucial to decouple the system from individuals
  • users and groups play roles but users and/or groups may change
  • delegation allows the permanent or temporary assignment of roles
  • without delegation many workflow systems have to deal with zombies

Management of User Data

• Integration with existing user information is crucial
  • critical data should never be managed redundantly
  • basic user information should be a basic IT service
• Many CMS allow to combine external data and internal configuration
  • keeping this combination consistent is not a trivial task
  • managing basic organizational information should be a central IT task

Workflows

• Publishing usually is a well-defined process
  • quality control is a very important reason for workflows
  • liability issues often have to be addressed by specialists
• CMS support for workflows is one of the main tasks
  • even wikis often move to a workflow with certain roles and rights
  • in large settings, managing workflows out-of-band is impossible

Workflow Example

Content

• Content may undergo changes during workflow stages
  • sometimes these changes do not change the content's structure
  • sometimes different schemas are used for different workflow stages
• Granularity of access rights and control can be important
  • for complex content, the document should not be locked if somebody makes small changes
  • managing changes with concurrent access is hard and often not supported
• Managing schema variations is not well-supported by basic Web technologies
  • some systems provide support for maintaining various variants of a schema
  • in most cases, it is up to the schema authors to guarantee consistency

Document Models

• XML often is used as the foundation for CMS content
  • XML is only a language for describing languages
  • without semantics and associated CMS functions, XML is not very useful
• CMS often have their own vocabulary for adding CMS support
  • depending on the access granularity, these are document-wide or subtree-based
  • the complexity of the vocabulary depends on the CMS data model complexity
• The document model is sometimes wrapped in CMS-specific structures
  • metadata for controlling access or workflow functions
  • metadata for describing the content for reuse and repurposing

DocBook

<!DOCTYPE book PUBLIC "-//OASIS//DTD DocBook XML V4.1.2//EN" "http://www.oasis-open.org/docbook/xml/4.1.2/docbookx.dtd">
<book>
  <bookinfo>
    <title>An Example Book</title>
    <author>
      <firstname>Norman</firstname>
      <surname>Walsh</surname>
    </author>
    <copyright>
      <year>2001</year><holder>Sun Microsystems, Inc.</holder>
    </copyright>
  </bookinfo>
  <preface><title>Introduction</title>
    <para>...</para>
  </preface>
  <chapter><title>The First Chapter</title>
    <para>...</para>
  </chapter>
  <!-- ... -->
  <appendix><title>An Appendix</title>
    <para>...</para>
  </appendix>
</book>

Darwin Information Typing Architecture (DITA)

• Topic orientation as an attempt to formalize classification
  • discrete units of information covering a specific subject with a specific intent
• Topic granularity for capturing different facets
  • self-contained topics combine with other topics into information sets
• Strong typing for tight control over document structures
  • DTDs and schemas guarantee that DITA types follow identical information structures
• Specialization as a way to add more refined classes
  • architecture for extending basic types to new types adapted for a particular use within an information set
• Common base class which is the root of the topic hierarchy
  • top-level generic base type provides fallback for all types

Old School

Declarative Web 2.0

Server-Side Annotation Service (Old School)

Server-Side Annotation Service (Declarative Web 2.0)

Client-Side Annotation Linkbase

Adding Annotations

• Adding annotations needs interaction beyond display
  • with the extension link creation could be supported as well
  • the extension would allow to create any XLink for a syllabus page
  • with a linkbase schema the extension could provide templates
• Adding annotations could be implemented in the application
  • the pipelines of the underlying CMS would have to be changed
  • storage could be managed locally or in a linkbase

CMS are Everywhere

• Not every CMS is called a CMS
• CMS functionality is used in many software products
• CMS need to integrate into existing IT environments
• Linkbases are special-purpose CMS