• integrate well into AsTMa*
  • retrieval and content construction processor
    • allows to phrase larger, more complex queries
    • better opportunities for global optimization, than if application would try it
    • means that the language needs a minimum of processing capabilities (if, loop, exceptions)
  • be embeddable into web application servers
    • normal retrieval language like SQL, XQuery (return simple values)
    • XML application servers, XML pipelines, taglibs (return XML fragments)
  • also as stand-alone ontology transformator
    • like XSLT for XML, transform maps between two different ontologies (return TM content)
  • allow to use typing
    • typing as in low-level types: AsTMa? does not do this (yet)
    • typing as in "using ontological information"
  • keep expressitivity of the language down, formal tractabillity