The definition of OWL, the ontology language underlying the Semantic Web, is based on formal representation methods. This provides benefits, in that tools have a firm definition of what they are supposed to do, but can have problems, due to difficulty or expense of building tools or mismatch with needs. The panel will discuss whether the general idea of designing standard Semantic Web languages with steadily increasing power (e.g., the progression from RDF to RDFS to OWL to OWL 2 to …) all based on formal methods is the right way to support the Semantic Web. What level of expressive power does the Semantic Web need? How should standard Semantic Web languages be designed? Does the Semantic Web even need formality?