This feature is especially useful for RDF. In the RDF case we will have one ontology per ontology file. But it is quite common to split large ontology files into multiple smaller files. We could improve performance significantly if we would put the content of all of these files into a single ontology:
load "rdf1.ttl" into <m1>
load "rdf2.ttl" into <m1>
Then both files are put into the same module.