doteduguru

Four Design Patterns - or How I learned to stop worrying and love the semantic web
TPR / 11:45 am / Monday
Brian Panulla @bpanulla

Why new web?

The SW is INFRASTRUCTURE.

• a parallel information architecture for smarter apps
• web content, pages and sites DO NOT need to change to be made ready for the 3.0 Webs

Three things the semantic web is NOT (kinda)

• it is NOT semantic HTML (thats a collision of terminology … this is a different semantic)
  • Zeldman says WRITE semantic HTML (use lists, use container p, use headers properly, don’t use presentational markup, don’t use br for padding and no tables for non-tabular data)
• it’s NOT Warmed-over Artificial Intelligence
• it’s NOT magic

Who cares?

(showing XKCD comic #773 about things on the front page of university websites — vs thins people go to the site looking for).  Everything on the right is semantic stuff!

Semantics

• explicit meaning of symbols
  • words - usage, connotation
  • images - symbolism
• become really useful when shared
• semantics differs from SYNTAX
  • rules of how symbols (words, letters, pictures) can be arranged.
    • I love Sushi.
    • I (heart symbol) sushi.
    • I love Sushi, but not in that way.
    • And who or what is “Sushi”?

Technology Primer

• RDF: Resource Description Framework
  • RDF is fundamental knowledge representation
    • declares resources
    • specifies properties of resources
  • What is a resource?
    • EVERYTHING is a resource… even properties.
  • URIs
    • resources are identified by a Uniform Resource Identifier (URI)
      • https://www.psu.edu/owl/list.owl#CollegeOfIST
      • https://2010.highedweb.org/sessions.rdf#TPR9
    • can also be round in your XHTML DTD and html tag’s xmlns attribute
  • Triples
    • RDF is expressed as triples:
      • Example 1
        • subject (“Penn State”)
        • predicate (“is a”)
        • object (“University”)
      • Example 2
        • subject (“Brian Panulla”)
        • predicate (“presented at”)
        • object (“High Ed Web Conference”)
    • Triples to Graphs
      • see how data is related
• Web Ontologies
  • where schemas describe structure, ontologies describe meaning or intended use
  • OWL adds more expressiveness and many aspects of formal logic to RDF
• On Design Patterns
  • Inspired by book Design Patterns: Elements of Reusable Object-Oriented Software
  • Architectural patterns like MVC

LINKED DATA

• Linked Data:  Use Cases
  • Publish/Syndicate complete information sets
  • Embedded explicit semantics, unique identifiers
  • have minimal impact to other Web information publishing
  • may be static or dynamically generated
  • EXAMPLES
    • GeoNames (geonames.org)
    • DBPedia (dbpedia.org)
    • Open Rubrics
  • The Linked Data Cloud
    • https://richard.evganiak.de/2007_10_lod (not sure on this URL …. )
  • Weaknesses
    • access control can be difficult for sensitive or confidential
    • damaging to proprietary intellectual property
  • Alternatives
    • serialized data: XML, JSON, CSV (common, structural not semantic)
    • microdata

MICRODATA

• Best Buy and Overstock have started tagging their products in RDFa
• Weaknesses
  • XHTML required for RDFa validation
    • new DOCTYPE
    • XML MIME type
• Alternatives
  • microformats (not discoverable)
  • HTML5 Data attributes - application-level semantics only
  • HTML5 Microdata - easier to read, more verbose elements

CONSUMER

• Manipulate linked data in a rich client
• generate multiple presentational forms
• i.e. Open Rubric Builder
• Weaknesses
  • processing/bandwidth resources constrained
  • client libraries still maturing
  • constraints of accessibility requirements
• Alternatives
  • AJAX, AHAH, JSON

SEMANTIC WEB APPLICATIONS

• Domains where schemas/models change rapidly or data is sparse
• Semantics of relational model inadequates (e.g. inferencing, inheritance)
• Domains emphasizing relationships
  • social networks
  • taxonomies
• Triple Stores
  • database tuned for graphs of RDF statements
  • sometimes a Quad Store (provenance)
  • Popular - Jena, AllegroGraph, Virtuoso
• Weaknesses
  • bad for opaque objects with few relationships
  • large sets of homogenous objects
• Alternatives
  • RDBMS
  • NoSQL (document dbs, key/value stores, graph databases)

SPARQL

• query language for RDF docs
• grabs all subjects, predicates and objects from RDF document
• queries are similar to mysql

Higher Education Use Cases

• Degree Requirements based on class listings
• Talk to Brian about other use cases … we ran out of time!!