Playing around with the RVF recently has raised some questions for me. (Mostly around Modelling, so MAG seemed like the best place to post).
(The RVF defaults to validating the MRCM against the “stated”. (Presumably this now means axioms).I’d have thought the inferred form is really what’s important? Otherwise problematic inferences aren't revealed)
Running the RVF against the inferred form I encountered many fails like
The MRCM attribute cardinality is [0..1] for 840562008 |Has device characteristic (attribute)| but found more than 1
e.g. : 395441002|Nelaton antimicrobial-coated urethral catheter (physical object)|
(it has 1 "Has device characteristic" stated, but 2 in the inferred view)
The MRCM does say this attribute has a cardinality of “0..1” and the concept has two properties, so the “failure” is valid…
However, the MRCM seems wrong…
It seems entirely reasonable to state that a device can have more than 1 characteristic (sterile AND coated)
(Like done with this concept 716644008|Sterile polymer coated metal esophageal stent (physical object)|)
And when a concept has more than 1 stated parent, it’s entirely possible for it to infer properties from both parents..
Then I look at the HRCM, DOM 260787004 Physical object - HRCM20210131 - SNOMED Confluence (ihtsdotools.org)
and I notice it talks about 3 views..
- General View
- Author View
- Expression View
What’s the difference between these three views? (I couldn’t see anything obvious)
I always thought, the concept model was the concept model.
I'm not sure what's going on, but wondered if the current drive towards "authoring templates" might be having some influence over the concept model?
(My question is not specifically about this attribute, but the role of the MRCM in general)
Contributors (0)
Comments