110 likes | 222 Views
Prolearn-WP1/4/5 12/2-2004. Proposal for an Abstract Model Query Language. Ambjörn Naeve Mikael Nilsson The Knowledge Management Research group Centre for user-oriented IT Design Royal Institute of Technology. http:// kmr.nada.kth.se. Ongoing relevant work.
E N D
Prolearn-WP1/4/5 12/2-2004 Proposal for an Abstract Model Query Language Ambjörn NaeveMikael Nilsson The Knowledge Management Research group Centre for user-oriented IT Design Royal Institute of Technology http:// kmr.nada.kth.se
Ongoing relevant work • Dublin Core Abstract Model (Powell, Nilsson, Naeve) • Common Model for DC and LOM (Powell, Duval, Nilsson, Naeve) • Edutella Query Language (QEL) (Siberski, Nilsson) • Simple Query Interface (SQI) (Simon, Duval)
Simple Query Interface • simple webservices interface to talk to a repository • works through executeQuery(String query) and some session handling around it • plans for several query languages, e.g. • QEL (Edutella, semantic search) • Xquery (syntax-seach) • free-text (checking each field, implem. dependent)
Simple Query Interface • ADVANTAGES: • a fast way to reach an interface where ”operators” can start talking to each other (semantic co-existence) • DISADVANTAGES: • no given paradigm (”paradigm co-existence”) • no common idea of how metadata is structured • (no abstract model ==> no semantic co-operation)
Edutella QEL Only one QEL in a given paradigm (RDF) • ADVANTAGES: • a general QEL ==> all sorts of data can be asked for • the co-existence can be used semantically, i.e. the common grounds can be used to build more advanced bridges between different systems • DISADVANTAGES: • the choice of RDF as a common paradigm imposes extra demands on the repository
Abstract Model Query Language (AMQL) • an abstract query language based on a high-level paradigm similar to the one represented by the DC abstract model • would have primitives of type: • ”give me all resources whose property [X] have the value [Y]” • where values could be assigned to X and Y
Abstract Model Query Language (AMQL) • ADVANTAGES: • not bound to - but compatible with - e.g. RDF and XML • could be translated to different paradigms, such as XML, RDF, database, … • provides a common basis for different metadata standards to plug into • DISADVANTAGES: • no one knows if it could be done, but the working group has the appropriate knowledge to find out!
Levels of implementation SQI Edutella AMQL Query Language API Physical architecture
Development strategy • provide AMQL as a query langage for SQI • add a physical architecture (e.g. query-hubs) • This would result in something that can match Edutella, and which would have the following • ADVANTAGES: • would not depend on a given paradigm (e.g. RDF) • could bridge the threatening gulfs between different paradigms (especially XML and RDF)