Scalable connection of models
Viewing those to be connected as big composite events so that analysis at the top layer can be done (and preserved) then the composites run on their own --- really efficient trace generation by recognizing those part of the schema that are logically independent of the others and so trace-generation can be isolated to preserve that which is already done to that which needs to change based on only the affected parts of the logic. On really big models this will help scaling. But also allows the building of the "drill-down" user interfaces to that composites can be the final node in a directed graph or expanded to include all of their interactions. COORDINATE statements and other constraints based on events within the composite event would then be simply applied to the top composite event as if it were atomic. Should be done in the RIGAL language.