50 likes | 171 Views
Report on the SCHAC initiative TF-EMC2 Poznan, 5 June 2005. SCHAC The beginning. A Committee inside TF-EMC2 to define and promote a common schema Contributing attributes and expertise Avoid reiventing the wheel in several shapes Build an initial kernel from already existing local attributes
E N D
Report on the SCHAC initiative TF-EMC2 Poznan, 5 June 2005
SCHACThe beginning • A Committee inside TF-EMC2 to define and promote a common schema • Contributing attributes and expertise • Avoid reiventing the wheel in several shapes • Build an initial kernel from already existing local attributes • Agreement on syntax and semantics • Allocation to object classes • Make the kernel evolve via a collaborative approach • Propose a new attribute/class • Period of comments by the rest of the group • Different voting rules for • Classes • Attributes • MAY or MUST definitions
SCHACSCHema for ACademia • The results of the SCHAC group • Not only oriented to directories • Trying to build a neutral schema: LDAP, XML... • The main goal is interinstitutional data exchange • Not pretending to substitute any existing schema • Coding rules taking into account schema idiosyncracies • Applicable to different domains • eduRoam, GEANT2, Grids • ECTS • Cotswolds Group, LoF • ECTS
Current status • RC1 of individual attribute definitions sent to the group • Attribute names, syntax and semantics • Extension of eduPerson (and naturalPerson?) • Based on the synchornization of attributes in schemas from • Croatia (hrEdu) • Finland (funetEdu) • France (supAnn) • Norway/Sweden (norEdu) • Poland (plEdu) • Spain (iris) • Switzerland (swissEdu)
Open issues • The naturalPerson question • The URN dilemma • Whether to use URNs for encoding • Relationships with other entities (specially if they can be multiple) • Values with local properties (instead of agreed controlled vocabularies) • Pros • Can act as controlled vocabularies • Resolvers can provide a level of additional flexibility • Delegation made them flexible with respect to local requirements • Better evolution with respect to versioning • Cons • Require careful namespace allocation • Need for parsers/resolvers/registers/... • Risk of partitioning the semantic space