First, these types have "fraction" parts, which probably should be unsigned, not signed.
Second, these types really want to be converted into an xs:dateTime type. But unfortunately, the DFDL libarary of functions does not provide rich enough functions to do this. A Daffodil pluggable User Defined Function (UDF) (or a handful of them) could be defined, and become part of this schema to enable conversion of these integer + fraction fields into a real xs:dateTime.
(These functions could be proposed for inclusion as Daffodil extensions to DFDL, and for inclusion in a future DFDL 2.0 language spec.)
(Currently, the only thing that constructs a xs:dateTime in an expression takes a string that looks like an ISO 8859 date+time. We need constructors that work in seconds and fractions of a second, with a way to specify the precision of the fractions of a second part.
Without this there is effectively no way to provide sensible range constraints on the elements that have these types.
Related DFDL Spec issue:
OpenGridForum/DFDL#8 (comment)
First, these types have "fraction" parts, which probably should be unsigned, not signed.
Second, these types really want to be converted into an xs:dateTime type. But unfortunately, the DFDL libarary of functions does not provide rich enough functions to do this. A Daffodil pluggable User Defined Function (UDF) (or a handful of them) could be defined, and become part of this schema to enable conversion of these integer + fraction fields into a real xs:dateTime.
(These functions could be proposed for inclusion as Daffodil extensions to DFDL, and for inclusion in a future DFDL 2.0 language spec.)
(Currently, the only thing that constructs a xs:dateTime in an expression takes a string that looks like an ISO 8859 date+time. We need constructors that work in seconds and fractions of a second, with a way to specify the precision of the fractions of a second part.
Without this there is effectively no way to provide sensible range constraints on the elements that have these types.
Related DFDL Spec issue:
OpenGridForum/DFDL#8 (comment)