Connecting Connexions: Organizing and Integrating Open

Learning Content with Topic Maps

Darina Dicheva

Lars Johnsen

Goal

Connexions: a repository of open educational resources sharing and reuse of learning modules under a Creative

Commons license

Modular design modules become “information islands” ways to organize and integrate modules’s content needed

Goal: organize CNXs open content using Topic Maps improved navigation more extensive reuse

Connexions

Connexions (cont’d)

Connexions (cont’d)

Connexions (cont’d)

Connexions (cont’d)

Connecting Connexions Content Our approach:

Extract content and metadata from Connexions modules and map them onto topic maps that can be used as a skeleton for further processing, enrichment and merging of subject-oriented learning content

Encode mappings as XSLT transformations Merge extracted topic maps related to a subject of

interest

CNXML Learning modules published as “structured XHTML”

contain mark-up reflecting their underlying XML structure <div> and <span> elements carrying a class attribute signal a certain

underlying category, e.g. <span class=”term”>…</span>. Each module is realized as two web pages, one containing the

actual learning content and one conveying relevant metadata CNXML comprises tags to mark up text elements in

modules: structural tags (section, paragraph, link and so on) semantic tags (example, definition, meaning, etc.) presentational tags (like emphasis) CNXML schema also incorporates elements to mark up different

media types, internal/external links, common metadata categories (date, version, keywords, etc).

CNXs-TMs Mapping with XSLT

Relatively easy, due to the consistent XHTML representation of the modules

Involves decisions on what data elements to search for, what TM constructs to draw upon and what topic structures to build

Distinguishes between “addressable” subjects, such as modules, sections or multimedia objects and “non-addressable” subjects, such as authors and topics represented by terms or keywords Topics are given subject identity (in TM terms): endowed

with more specific identity markers such as subject identifiers (for non-addressable subjects) or subject locators (for addressable subjects)

TM4L Connexions Plugin

The result of a topic map extraction from the Connexions module “Guitars”

The merged topic maps

The merged topic maps

The merged topic map and its visualization

CNXs Plugin Implementation

Implementation XSLT style sheets XSLT processor integrated in TM4L (Apache ‘s Xalan-Java

Version 2.7.1) The XSLT processor is an important extension of

the overall functionality of TM4L since it supports general XML transformations experienced TM4L authors can import or write their own

XSLT transformations for topic map generation or for rendering a topic map in a certain way

Work on constructing a web crawler capable of harvesting web pages linked to the page under transformation is currently underway

Thank you