Closed jakebeal closed 2 years ago
Assigning this to library maintainers, since we believe this change can be made without an SEP if all of the library maintainers agree.
@udp and @tcmitchell have indicated assent; @goksel , I assume you're OK with it since you've already got a branch in the ontology incorporating the change. Can you please confirm?
A number of files in SBOLTestSuite/SBOL3 will need to be updated to implement this change:
@goksel can you update these files on a branch (feature/sbolcomposition
as in sbol-owl3?) so that we can coordinate development of this change?
I found these issues after regenerating the SHACL rules using https://github.com/SynBioDex/sbol-shacl. I pushed a new feature/sbolcomposition
branch that contains this change in case others want to try out the updated SHACL rules.
@goksel has affirmed as well, so I am going to set up a pull request for this change in the spec.
I believe this should have been closed already when #463 was merged.
The
hasFeature
property ofComponentReference
conflicts in type with thehasFeature
property ofComponent
, because inComponentReference
this is a pointer while inComponent
it is a child object. This causes difficulty for defining the ontology, because the property has contextual meaning.We propose that
hasFeature
was never actually a good name for theComponentReference
property in the first place, sinceComponentReference
's other property,inChildOf
is also aFeature
(subclass). We thus propose to resolve the type conflict by renaming ComponentReference torefersTo
, which is a better description of its meaning.