Uploaded image for project: 'FHIR Specification Feedback'
  1. FHIR Specification Feedback
  2. FHIR-33545

Notational conventions useful and confusing

    XMLWordPrintableJSON

Details

    • Icon: Change Request Change Request
    • Resolution: Persuasive with Modification
    • Icon: High High
    • Shorthand (FHIR)
    • 1.0.0
    • FHIR Infrastructure
    • Normative
    • Language Reference
    • 3.1
    • Hide

      The commenter is correct, it is tricky to create a meta-language to describe the syntax of a language. We included the formal description of FSH in ANTLR, but it is not helpful to the majority of readers. In addition, ANTLR does not capture all the nuances of the language, only the syntax. We point this out when we introduce notation used to describe FSH.

      Rather than discarding the meta-notation that was rather difficult to come up with, and has proven valuable in describing the language with all its nuances, we propose to review the points made by the commenter and see if there are specific clarifications that can be made. 

      Show
      The commenter is correct, it is tricky to create a meta-language to describe the syntax of a language. We included the formal description of FSH in ANTLR, but it is not helpful to the majority of readers. In addition, ANTLR does not capture all the nuances of the language, only the syntax. We point this out when we introduce notation used to describe FSH. Rather than discarding the meta-notation that was rather difficult to come up with, and has proven valuable in describing the language with all its nuances, we propose to review the points made by the commenter and see if there are specific clarifications that can be made. 
    • Mark Kramer / Rick Geimer : 16 - 0 - 1
    • Correction
    • Non-substantive
    • Yes

    Description

      The content in this section is useful and confusing at the same time.

      The section is trying to come to a shorthand definition for defining shorthand that stands somewhere in the middle between formal ANTLTR and text. Although I do see the value of the notation in the specification, as it is more accessible than when using ANTLR, it does leave more room for interpretation. It also introduces another layer in which inconsistencies can occur.

      Also some of the concepts introduced in this section do not appear in the grammar (e.g. CodeableConcept, datatype) or are using a different name (e.g. decimal vs NUMBER). Sometimes the grammar uses the same keyword for something else (e.g. Resource in this section relates to a resource name, in the grammar it is a resource definition).

      Resolving this is tricky, but seeing the importance of this semi-language in the spec, I think this has be done.

      One approach could be to separate the notational elements from the definitions and list or refer to the exact definitions for those elements that relate to the formal language specification; expressing each syntax term in ANTLR using the formal ANTLR definition.

      Attachments

        Activity

          People

            Unassigned Unassigned
            bvdh Bas van den Heuvel
            Bas van den Heuvel
            Watchers:
            2 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved: