Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Introduction to Application Profiles


Published on

Presented January 18, 2010 to the ALCTS Committee on Cataloging: Description and Access (CC:DA) as an introduction to RDF data, and application profiles. Presenters were Jon Phipps, Karen Coyle and Diane Hillmann.

Published in: Technology, Education
  • Be the first to comment

Introduction to Application Profiles

  1. 1. Jon Phipps: Overview<br />Karen Coyle: Step-by-Step<br />Diane Hillmann: Context<br />Application Profiles<br />
  2. 2. What&apos;s an Application Profile?<br /> <br />
  3. 3. What&apos;s an Application Profile?<br />It&apos;s a document<br />
  4. 4. What&apos;s an Application Profile?<br />It&apos;s a document of an agreement<br />
  5. 5. What&apos;s an Application Profile?<br />It&apos;s a document of an agreement on a model<br />
  6. 6. What&apos;s an Application Profile?<br />It&apos;s a document of an agreement on a model of our stuff in the world<br />
  7. 7. What&apos;s an Application Profile?<br />It&apos;s a document of an agreement on a model of how we describe our things in our world<br />
  8. 8. What&apos;s an Application Profile?<br />It&apos;s a document of an agreement on a model of how we describe ourthings in our world (domain) in the context of the global web of data<br />
  9. 9. Things?<br /> <br />
  10. 10. Things?<br />have a formal definition...<br />
  11. 11. Things?<br />have a formal definition...<br />Every individual in the OWLworld is a member of the classowl:Thing.<br />
  12. 12. OWL?<br /> <br />
  13. 13. OWL?<br />Web Ontology Language<br />
  14. 14. OWL?<br />Web Ontology Language<br />A language that can be used to formalize a domain by defining classes, the relations between them, and properties of those classes<br />
  15. 15. OWL<br />Web Ontology Language<br />can define the semantics of an Application Profile<br />
  16. 16. Semantics?<br />
  17. 17. Semantics?<br />What we mean when we define a class called &apos;book&apos; and describe it with a property called &apos;title’.<br />
  18. 18. Semantics?<br />What we mean when we define a class called &apos;book&apos; and describe it with a property called &apos;title&apos;.<br />The &apos;Semantic Web&apos; is a web of meaning that uses the RDF model<br />
  19. 19. RDF?<br />Resource Description Framework<br />
  20. 20. RDF?<br />Resource Description Framework<br />
  21. 21. RDF?<br />Resource Description Framework<br />“is a framework for representing information in the Web”<br />
  22. 22. RDF?<br />“has a simple data model that is easy for applications to process and manipulate.”<br />
  23. 23. RDF?<br />“has a formal semantics which provides a dependable basis for reasoning about the meaning of an RDF expression.”<br />
  24. 24. RDF?<br />“URI references are used for naming all … things in RDF.”<br />
  25. 25. RDF?<br />“is an open-world framework that allows anyone to make statements about any resource.”<br />
  26. 26. RDF?<br />“The underlying structure of any expression in RDF is a collection of triples”<br />
  27. 27. Triples?<br />consist of a subject, a predicate and an object. A set of triples is called an RDF graph<br />
  28. 28. Triples?<br />consist of a resource, a property and a value surrogate. <br />
  29. 29. Value Surrogate?<br />Not a value, but some thing that denotes the value<br />
  30. 30. Value Surrogate?<br />can be a Literal or a non-Literal<br />
  31. 31. Literal?<br />Can be a Plain Literal or a Typed Literal<br />
  32. 32. Plain Literal?<br />Is just a string with an optional (in this case totally unnecessary) language type<br />Plain Literal: <br />“Samuel Clemens”@en-US<br />
  33. 33. Typed Literal?<br />A string that must be interpreted<br />Typed Literal: <br />“27” ^^xsd:integer<br />
  34. 34. Typed Literal?<br />A string that must be interpreted<br />Typed Literal: <br />“27” ^^xsd:integer<br />denotes the number 27 <br />
  35. 35. Non-Literal?<br />A URI that refers to a resource<br />
  36. 36. Non-Literal?<br /><br />Identifies the skos:Concept labeled<br />“sounds”@en-US<br />In the skos:ConceptScheme identified by the URI<br /><br />
  37. 37. URI?<br />A resource identifier.<br />
  38. 38. URI?<br />A globally unique resource identifier.<br />“All URIs share the property that different persons or organizations can independently create them, and use them to identify things.”<br />
  39. 39. Predicate?<br />A URI<br />that identifies the property of the subjectof the triple<br />
  40. 40. Predicate?<br />“Since RDF uses URIs instead of words to name things in statements, RDF refers to a set of URIs (particularly a set intended for a specific purpose) as a vocabulary”<br />
  41. 41. Property?<br /> the property labeled<br />“Place of production”<br />in the vocabulary identified by<br />
  42. 42. Property?<br />A vocabulary can declare a property to be a subproperty of another property.<br />
  43. 43. Property?<br />A vocabulary can declare a property to be a subproperty of another property.This creates a formal relationship between the properties<br />
  44. 44. Can we talk about APs now?<br />Please?<br />
  45. 45. An AP defines Semantics<br /><ul><li>The Classes of resources your metadata is describing
  46. 46. The Vocabularies that you will use as properties to describe them</li></li></ul><li>Classes?<br />OMG<br />Please, don’t start that again<br />
  47. 47. An AP defines Syntax<br /><ul><li>Valid value ranges and datatypes for each property
  48. 48. Valid lists of values (controlled vocabularies) for properties
  49. 49. Cardinality of each property</li></li></ul><li>An AP defines Syntax<br />Dublin Core defines this validation profile for each property as a“Statement Template”<br />
  50. 50. An AP defines Syntax<br />A set of Statement Templates is a “Description Set”<br />
  51. 51. An AP defines…<br />An AP can describe multiple Description Sets.<br />
  52. 52. An AP defines…<br />An AP can describe multiple Description Sets.<br />The full set of Description Sets is a “Description Set Profile”<br />
  53. 53. Application Profiles<br />Step-by-Step<br />1/18/2010<br />52<br />CC:DA Application Profile Intro<br />
  54. 54. 1. Domain model<br />Person<br />WEMI<br />Topic<br />FRBR<br />1/18/2010<br />53<br />CC:DA Application Profile Intro<br />
  55. 55. 2. Determine elements<br />Work<br />Title<br />Format<br />etc<br />1/18/2010<br />54<br />CC:DA Application Profile Intro<br />
  56. 56. 3. Identify vocabularies<br />1/18/2010<br />55<br />CC:DA Application Profile Intro<br />
  57. 57. Identify vocabularies<br />1/18/2010<br />56<br />CC:DA Application Profile Intro<br />
  58. 58. Identify vocabularies<br />1/18/2010<br />57<br />CC:DA Application Profile Intro<br />
  59. 59. Vocabulary do&apos;s & don&apos;ts<br />Do not select elements based on their names or labels<br />Do select elements based on their definitions<br />Do pay attention to what values can be used<br />Don&apos;t think that you can select an element that doesn&apos;t quite match your need, and use it anyway<br />Do think: INTEROPERABILITY<br />1/18/2010<br />58<br />CC:DA Application Profile Intro<br />
  60. 60. One vocabulary<br />Vocabulary<br />AP<br />1/18/2010<br />59<br />CC:DA Application Profile Intro<br />
  61. 61. More than one vocabulary<br />Vocabulary A<br />Vocabulary B<br />AP<br />1/18/2010<br />60<br />CC:DA Application Profile Intro<br />
  62. 62. Rolling your own<br />Vocabulary<br />AP<br />1/18/2010<br />61<br />CC:DA Application Profile Intro<br />
  63. 63. Rolling your own<br />Vocabulary<br />All elments must be defined outside of the AP. <br />AP<br />1/18/2010<br />62<br />CC:DA Application Profile Intro<br />
  64. 64. &quot;Constraints&quot;<br />Mandatory/optional<br />Repeatability<br />Values (cannot conflict with defined element)‏<br />Free text<br />Controlled list of values<br />Formatted text (e.g. dates)‏<br />1/18/2010<br />63<br />CC:DA Application Profile Intro<br />
  65. 65. What is the impact of all this on our world?<br />The Context for Application Profiles <br />
  66. 66. In the world we knew, interoperability was ensured by “compliance with standards”<br />All of used the same ones in a closed world<br />Data created by humans under strict guidelines<br />In an open world, interoperablity depends on:<br />Technologies that reach beyond one community<br />Data built in a variety of ways by people with different ideas of the world<br />Machines that act broadly based on human oversight<br />Understanding Interoperability<br />1/18/2010<br />65<br />CC:DA Application Profile Intro<br />
  67. 67. We’ve long accepted the limits of requiring upfront consensus to ensure interoperability<br />In a world of APs we can specialize beyond the core of generally useful data<br />We don’t need humans to ‘dumb down’ specialist data to enable sharing and interoperability<br />Machines can invoke relationships to generalize specialist data when necessary, without removing the value of extended specialized data for specialists<br />Why This Approach Instead?<br />1/18/2010<br />66<br />CC:DA Application Profile Intro<br />
  68. 68. We can’t afford to “go it alone”<br />We can’t afford to ignore the world of data outside libraries<br />We can’t afford to create all our data with humans<br />We can’t afford NOT to rethink how we operate<br />The Value of an Open World<br />1/18/2010<br />67<br />CC:DA Application Profile Intro<br />
  69. 69. It’s often free and easily available<br />It’s ‘good enough’ (our stuff isn’t perfect either)<br />It takes us where we can’t go with our current data<br />It’s maintained by someone else <br />We can choose to use data or not, APs allow us to document that use, automate the process, and expose the data to others<br />What’s Out There? Data!<br />1/18/2010<br />68<br />CC:DA Application Profile Intro<br />
  70. 70. Records can be aggregated from statements when we need them<br />Statement-based data can be managed and improved more easily than record-based data<br />Statement-based data can carry provenance for each statement, allowing quality decisions to be made at a more granular level<br />Changing Our Data Management Ideas<br />1/18/2010<br />69<br />CC:DA Application Profile Intro<br />
  71. 71. Getting From Here to There<br />1/18/2010<br />70<br />CC:DA Application Profile Intro<br />
  72. 72. The RDA Vocabularies<br />The principles of extension inherent in the RDF Vocabulary standards used<br />Our experience in building and using data<br />Using What We Already Know<br />1/18/2010<br />71<br />CC:DA Application Profile Intro<br />
  73. 73. Proliferating our ideas and experience with bibliographic data to the broader Web world<br />Using newer technology to achieve more efficiency, transparency, and functionality<br />If retrenchment is the only answer, the end point is zero<br />Saving our precious human resources to think, evaluate, ensure quality, and innovate<br />To Build Ourselves a New Future<br />1/18/2010<br />72<br />CC:DA Application Profile Intro<br />
  74. 74. We can map it in a variety of ways for a variety of uses<br />We can still use MARC as a [lossy] exchange format as long as we need it<br />It offers insufficient flexibility as the basis for a new data world inter-connected to the Web<br />We can use our skills and our understanding of bibliographic description to lead the way forward<br />What About Our Legacy Data?<br />1/18/2010<br />73<br />CC:DA Application Profile Intro<br />
  75. 75. Specialist communities are already thinking about what they need that RDA doesn’t provide<br />Using the extensibility of RDF vocabularies allows them to choose from a number of options <br />Moving proposals through the RDA process<br />Extending the vocabularies through their community domain<br />Chosing the extension option reflects the reality that consensus has its limits, and specialist data may be better managed at a different level<br />Some Concrete Suggestions<br />1/18/2010<br />74<br />CC:DA Application Profile Intro<br />
  76. 76. With Application Profiles we can:<br />Document our decisions clearly<br />Measure compliance with our own intentions<br />Express our decisions in a machine-actionable way<br />Make connections with other data communities by re-using their data semantics<br />RDA expresses this ideal in its stated goals<br />What Do We Gain?<br />1/18/2010<br />75<br />CC:DA Application Profile Intro<br />
  77. 77. Less one-at-a-time creation and more data design, data improvement, data evaluation<br />Ability to look at our contrained resources and reduced budgets as the opportunity to reinvent ourselves<br />Rethinking Our Role<br />1/18/2010<br />76<br />CC:DA Application Profile Intro<br />
  78. 78. Guidelines for Dublin Core Application Profiles<br /><br />RDA Vocabularies<br /><br />Thanks! Links! Questions?<br />1/18/2010<br />77<br />CC:DA Application Profile Intro<br />