In my view UML and SysML are fluffy apps designed to entice unwary developers into purchasing software that they will then get locked into.
The tension has always been between doing something that is formally rigorous (such as an ontology and logic language ) versus adopting something that is convenient for users to use (such as UML and SysML). This is the world of expediency, where everyone wants to take shortcuts, with the end result that you have something that looks OK and didn't take to long to produce, but that is ultimately useless. I think that is the refrain of the crowd from your Google+ post.
This is a NASA JPL presentation from a few years ago that I refer back to:
> [Semantically-Rigorous Systems Engineering Using SysML and OWL](http://esaconferencebureau.com/docs/12c12_docs/0910-jenkins.pdf?sfvrsn=2)
OWL is the standard "web ontology language"
And more recent info here -- note the "ease of adoption" reference, which is a code word for expediency:
International Symposium on Leveraging Applications of Formal Methods (2016)
I personally do analysis by creating graphs on the fly from a logic language connected to a triple-store database. At some point everyone that works in this area has an epiphany and figures out that everything is connected by triples, and governed by predicate logic, and that is a rigorous way to grow a model.