Add support for Real values with DV_SCALE
Description
Attachments
CR fixes the problem in
relates to
Activity
Ian McNicoll November 2, 2020 at 4:41 PM
“The other consensus is if there is at least one Real in the custom 1.4 syntax, this is interpreted as a DV_SCALE, otherwise (only integer) it is assumed to be a DV_ORDINAL. We’ll do better in 2.0 with tuples etc.“
Agree - we need to respect that (for stupid reasons mostly) there will me a mix of integers ans reals in many cases.. but any real makes it a SCALE.
Sebastian Garde November 2, 2020 at 3:45 PMEdited
OK thanks Pieter - the same section does not exist in 1.4…so strictly speaking this is not correct in 1.4 land then as it stands.
However if I understand the commenters correctly, it seems to be the consensus here that Integers should be supported in a DV_SCALE in the same way? I think this consensus (if it is one) should be at least written down here if that is the case, so that parsers etc. deal with this in the same way (i.e. enable or reject an integer).
The other consensus is if there is at least one Real in the custom 1.4 syntax, this is interpreted as a DV_SCALE, otherwise (only integer) it is assumed to be a DV_ORDINAL. We’ll do better in 2.0 with tuples etc.
See this archetype with a few examples…
If you have any objections to these clarifications, please raise them now so that we and others can go and implement this for 1.4 in a consistent way.
Pieter Bos November 2, 2020 at 12:46 PM
I just checked, and it is specified in https://specifications.openehr.org/releases/AM/latest/AOM2.html#_rm_type_substitutions . Also note the two paragraphs before that. Maybe an example could help there?
Sebastian Garde October 30, 2020 at 5:19 PM
Yes - sorry I did not mean to imply that it is incorrect; just that it is confusing (to me anyway).
A note with a pointer to whereever it is documented what ADLW and Archie do, or just a note would probably be helpful….even more so if this is behaviour has been changed.
Pieter Bos October 30, 2020 at 2:21 PMEdited
The constraints are still of type real/C_REAL, so that documentation is correct.
There is simply an RM specific type substitution layer added later, saying that you can constrain a Real number in the OpenEHR RM with a C_INTEGER. It is also used for the terminology code difference, where the RM model is different than the AOM model.
At least, that’s part of the ADL workbench and Archie in the same way, I’m not sure if and where it is in the documentation.
See SPECPR-44.