Notes from 52nd International STC Conference
Seattle, Washington, May 8-12, 2005

Introduction to Metadata and RDF

Neil E. Perlin
Hyper/Word Services

Perlin is a recognized expert in "bleeding-edge" technologies. Each year at the STC international conference, he sponsors a series of technical panel discussions on the latest advances in technology.

Session Description: This session explained how to find specific content in a sea of topics by using metadata – "information about information."

• "Meta" is the Greek prefix for "beyond." It refers to data beyond (or transcending) other data, or information about information. It refers to information describing a web page or other content to help us find and use it. It's a high-tech idea that's existed for years as the mundane library card.
• Online-oriented metadata:
   
• MPEG-7: for A/V objects like audio, images, music, and video.
• XMP: extensible metadata platform: RDF-based standard from Adobe
• Dublin Core: for web resources
• RDF: resource description framework (W3C metadata standard)
• Types of metadata:
   
• Descriptive: help locate material
• Processing: what to do with it
• Administrative: manage it
• Metadata concepts:
   
• Schema: a set of metadata elements for a specific industry or purpose
• Controlled vocabulary: a required or recommended set of keywords
• Taxonomy: hierarchical classification system (but often subjective, with no rigor to the classification)
• Ontology: a constrained taxonomy, with more rigorously described item properties. It describes relationships between properties in such a way that the computer can draw logical inferences between items
• What does metadata do for us? Why should we care?
   
• Find files or content chunks
• Process those files or chunks
• Manage those files or chunks
• Support CMS use for single sourcing
• Enable the Semantic Web
• Beneath the hype:
   
• Metatags must be accurate, consistent, and honest
• Metatagging requires understanding of content
• Metatagging requires enough time to do it right
• HTML's <meta> lets us add keywords to HTML files. Web designers use them to try to maximize search engine hits on sites
• Metadata online: we've actually been creating and using metadata online for year, without calling it that: Windows Explorer, Word, RoboHelp.
• Creating metadata in Word:
   
• Basic metadata: select File/Properties
• Custom metadata: select File/Properties, Custom tab
• To search for a file: select File/Open and the Find function<
• Metadata and single sourcing:
   
• Open standards are the next step after tool-specific metadata. They can be a gamble in that they must apply across multiple tools and must be adopted by multiple tools. Many open standards exist but none yet dominates for documentation.
• Two examples of open standards are Dublin Core and RDF.
• Dublin Core:
   
• Born at a 1995 conference in Dublin, OH
• It is a core of 15 elements for describing web or other types of resources:
   
• Title
• Creator
• Subject
• Description
• Publisher
• Date
• Language
• Contributor
• Source
• Format
• Type
• Resource identifier
• Relation
• Coverage
• Rights
• It's simpler to understand than RDF and it can be coded in HTML and XHTML.
• Dublin Core syntax is unlike XML or other languages (all elements are optional and there is no required element sequence).
• RDF is a highly structured, XML-based web metadata standard that can be used to make assertions about web resources, as opposed to just identifying resources. This allows "intelligent" searches and single-source format conversions.
• Assertions are 3-part statements called triples (e.g., "This presentation is given by Neil," where the subject is "this presentation," the predicate is "is given by," and the object is "Neil." (NOTE: This is not grammatically correct; technically "this presentation" is the subject, "is" is the linking verb, and "given by Neil" is the complement. But the point is the basic syntactical pattern of S-V-O or S-V-C.)
• RDF is not yet extensively used in documentation, but it has great potential for creating accessible web sites that follow W3C standards and for creating Semantic Web projects. It is likely to come into much greater use over the next few years.
• The likely progression will be increasing use of programming in RDF, at the code level, followed by development of GUI tools such as those that automated the use of HTML.
• SWAP. Semantic Web Accessibility Platform. Using RDF metadata, you can take online material, mark it up automatically through a wizard, and run it through several processes to bring it out in several "layers" including 508-compliant versions (without altering the original). Can also accommodate PDA formats, etc., in single-sourcing fashion.
• Israeli company that has made break-through progress in automating 508-compliant web site accessibility: www.ubaccess.com
• Good procedures to get ready for the next wave (RDF, metadata, fully accessible web pages, Semantic Web): start developing good metadata habits.
   
• Fill out the Properties sheet or equivalent for any application you're using
• Start to add Dublin Core metadata to your HTML files, especially those in web-like formats like WebHelp or FlashHelp
• Start changing company practices and culture to support this technology trend.
• Summary of metadata and RDF:
   
• It's an esoteric technology whose time is now and is coming.
• There is no one dominant standard now.
• Today's metadata tools aren’t friendly.
• Expect to work in code, with many peculiarities, like with HTML in 1994, before the GUI tools become available