Lightweight Semantics

Download Report

Transcript Lightweight Semantics

Semantic Web Technologies
• Brief Readings Discussion
• Class work: Research topics and Project
discussion
• Research Presentation Topics assigned
• Building lightweight semantics
What are Lightweight Semantics?
• The bottom-up, user created, context-driven
set of markup conventions (not standards)
that can be used to provide associative
metadata in human readable formats
• Metadata
- For music
- For your Web browsing
- For your documents
• Microformats
• Emergent Semantics
Taking a Stand on the Sem Web
• Can the semantic Web scale?
- Can there be enough agreement for true sharing?
- How much can be predicted based on current use?
• Should the semantic Web be like MARC
records?
- What about MARC-up consistency?
• Training & standards
- Who is the LoC or OCLC in the Sem Web?
- Are Web Services & APIs like Z39.50?
• Semantic SPAM?
- Web of Trust
Stand for Lightweight Semantics?
• “any underlying assumptions about use are highly
situated”
- Does this imply that only experts should tag things?
• But the algorithms & heuristics are assembled by
experts right?
• How do you embed good semantics?
- At time of use?
- In the creation system?
• Is all this tagging at the cost of the user?
- Time to tag, time for finding
- But what if I organize my own data as well?
• Why can’t the SW experts be the same people as the
MARC experts?
Microformats
• “Designed for humans first and machines
second, microformats are a set of simple,
open data formats built upon existing and
widely adopted standards. Instead of throwing
away what works today, microformats intend
to solve simpler problems first by adapting to
current behaviors and usage patterns (e.g.
XHTML, blogging).”
• Let’s go to the site
Microfomats
• A combination of elemental formats that
work with compound formats
• Some are standards, some are becoming
standards through use
Lightweight Mark-Up
•
•
•
•
Markup for accessibility helps with the Semantic Web
Validating markup helps build the Semantic Web
“Semantic Overhead” during Web page creation
Use simple semantics for context
- Class
- Id
- Meta tags
• A set of tools to help
• This is an example of a paper to build from for a
research presentation
Emergent Semantics
• Ad hoc semantics
• Use known markup formats (XHTML) for
representing your structures
- Like using CSS to control display
- Easier to read than some semantic (meta) data
- “rel” values
• Easily transformed for more structured use?
- Working with APIs
• Eric Meyer’s presentation
Let’s use some microformats
• Build some lightweight markup to describe
your friendships
• The hard way:
- <a href="http://www.ischool.utexas.edu/~donturn"
rel="friend met colleague neighbor">Don
Turnbull</a>
• The easy way: http://gmpg.org/xfn/creator
Let’s talk about topics
• Discuss your topic ideas with your new XFN
friend
• What are the basic ideas that apply to the
Semantic Web, Web 2.0, Web Services, or
Lightweight Semantics?
• How would you show the ideas?
- examples
• What do the ideas build upon?
- research
Topic Selection
• Choose a topic (and corresponding week) to overview
• Topic Presentations should include:
-
Overview of the technology
Provide examples of the technology in use
Show how to build using the technology (examples)
A list of citations and readings that you drew from and for
extended reference
• Do not rely on wikipedia & blogs as your only sources
• Academic journal & conference papers
• Books (development or conceptual design)
• How can these Semantic Web technologies help
coordinate, discover, organize information and
knowledge?
• Your own point of view about the practicality &
promise of these tools & procedures
Current list of Topics
•
•
•
•
•
•
•
•
•
•
•
•
•
RDF
Metadata (e.g. Dublin Core, MediaRSS)
Ontology building (applications)
REST, XMLHttpRequest & AJAX
Greasemonkey
Javascript: Introduction
Javascript: Advanced
TagClouds
GIS, Maps & Mapping Mashups
XSLT
WordNet
Semantic Commerce
Trust