Currently editing RO requires a lot of tacit knowledge
SOP for making inverse relations
When to encode certain things as SWRL, and some as DL, and how not to run afowl of DL restrictions
eg has-component hack
rolification and interaction relations
relations defined by
a genus relation and either domain, range, or both
a property chain (e.g. overlaps = has-part o part-of)
intersection
union
etc
We have attempted to document some of this in the RO docs but these are textual and it is still hard for editors to know (a) the pattern to choose when adding axioms (b) how to know if they get it right
When discussing this issue #521 as an exemplar case, on the RO call 2 months ago we realized that it would be hard to adapt either dosdps or robot templates to this solution as we frequently need to coordinate sets of relations and sometimes helper classes in the same "record"
cmungall
commented
2 years ago
Currently editing RO requires a lot of tacit knowledge
We have attempted to document some of this in the RO docs but these are textual and it is still hard for editors to know (a) the pattern to choose when adding axioms (b) how to know if they get it right
slides
When discussing this issue #521 as an exemplar case, on the RO call 2 months ago we realized that it would be hard to adapt either dosdps or robot templates to this solution as we frequently need to coordinate sets of relations and sometimes helper classes in the same "record"