Arriving at a Standard Set of Categories

See DcSubject for a place to list different categories in use by Indymedia sites.

Problem statement

How can we create a standard set of categories between IMCs spread out over 6 continents in dozens of different languages?

Necessary features:

• Multi-lingual
• Unique reference to heavily overlapping categories
• Extendable for local topics
• Can easily add new terms that we didn't thing of at first
• Matches de-centralized nature of Indymedia; avoiding central repository?

Similar Problems

• Cataloging reference materials
• Finding file types (the MIME system)

Possible Approaches

• Use external source
• Generate exhaustive list and combine duplicated subjects
• Don't define exhaustive list -- just a way to reference which IMC defined the category (e.g., if UK defines a category for environmentalism, it would be uk.indymedia.org / Environmentalism)
• Allow multiple definitions of the same subjects, using different controlled vocabularies and the dc:subject scheme to separate them. This website gives a good example of this approach.

Starting Points

• Library of Congress Subject Headings (but not internationalized?)
• Dewey Decimal Classification
• NewsML/IPTC Newscodes -- 1300 subject terms translated into german, english, french and spanish.
• http://www.lub.lu.se/metadata/subject-help.html
• http://www.thegateway.org/about/documentation/metadataElements/subjectRefinements/subject-6-6-04

Examples

IPTC classifications for news articles

The broad categories: many, many subdivisions.

Example: LOC Classification Numbers

(without sub-headings)

Dewey Decimal

Universal Decimal Classification

First pass at an IMC Controlled Vocabulary

This approach uses the IPTC newscodes as a starting point. It moves protest (a subset of the 16s) to their own category and adds a few other new categories, including one for Indymedia. Categories with a ? are less likely to be used.

General approach: up to 99 broad categories, 01-99, with sub-categories. E.g., broad category: 01000000. Subcategory: 01100000. No need for "synthetic" classification approach since each article/feature can have multiple associated categories.

Goal: make a sensible hierarchy, without leaving out any important IMC categories. Secondary goal: make at least part of our hierarchy compatible with the IPTC newscodes.