Consistent use of canonical units and scientific notation would simplify machine interpretation. - 2018-May Core Norm Infrastructure #10

XMLWordPrintableJSON

    • Type: Change Request
    • Resolution: Persuasive
    • Priority: Medium
    • FHIR Core (FHIR)
    • STU3
    • FHIR Infrastructure
    • Datatypes
    • 2.23.0.1
    • Hide

      Motion: Indicate that we allow the exponential form in all formats. This means that in XML, a FHIR decimal might map to x:decimal or xs:double (as a union) and that implementations will need to populate and parse accordingly.

      One reason to do this is that the number of significant digits in a number like "1000" is unclear, whereas 1.00E3 is clear.

      Show
      Motion: Indicate that we allow the exponential form in all formats. This means that in XML, a FHIR decimal might map to x:decimal or xs:double (as a union) and that implementations will need to populate and parse accordingly. One reason to do this is that the number of significant digits in a number like "1000" is unclear, whereas 1.00E3 is clear.
    • Grahame Grieve/Jeff Danford: 5-0-0
    • Enhancement
    • Compatible, substantive
    • STU3

      Existing Wording: "decimals SHALL not use exponents, and leading 0 digits are not allowed?"

      "exponents are not allowed"

      Comment:

      The problem arises for large values with low precision. It is impossible to express these using the decimal type. The statement ?that large [?] values are extremely rare in medicine? is due to a human-centric representation of values; this is opposed to a machine-centric representation or scientific representation. The machine-centric representation would benefit from a more canonical representation. This would ease machine learning and artificial intelligence uses for the data by allowing for the representation with common units rather than having to convert between representation of values in various units selected specifically to maintain a number of precision digits . Consistent use of canonical units (SI units and derived units) and scientific notation would simplify machine interpretation.

      Specific uses of this include in the imaging domain for representing measurements and dosages. It would also be applicable for representing labs so avoid the confusion of selecting units of measure optimized for precision.

      Summary:

      Consistent use of canonical units and scientific notation would simplify machine interpretation.

            Assignee:
            Unassigned
            Reporter:
            Chris Melo
            Watchers:
            2 Start watching this issue

              Created:
              Updated:
              Resolved: