Pros and cons of this approach


  • It works

  • Time efficient

    • Even complicated compound languages can be defined within minutes

    • No conversions, no adjustments of original schemas

    • Rewriting the NVDL definition for a different compound language (different parent language, additional language involved or different contexts) in minutes

  • Less knowledge required

    • The previous NVDL script is composed only of 7 elements and the whole NVDL specification only has 10 more

    • NVDL script authors don't need to have any knowledge of the schema languages used for referenced grammars

  • Fully reusable

    • The schemas for involved languages can be fully reused as they are (even at their original location)

    • No need to download them, convert them or adjust them in any way

    • Whether schemas are written in W3C XML Schema, RELAX NG or Schematron, no matter NVDL can handle all of them in one validation process

    • They may be reused as they are for any compound language definition imaginable

  • Easy to read and understand

    • NVDL scripts are easy to read (just few elements)

    • They contain only information about the compound language (not the individual languages)


  • I don't see any significant cons

  • Few little issues are address in our extensions to the NVDL specification