Multiple Ranges and Domains
With respect to the third comment at  about unintended domain and range specifications I agree that the intended semantic very likely is that of an OR.
However, I would like to point out that any OR statement (more precisely disjunction) will push the rNews ontology into the most complex OWL 2 language fragment. In particular, the ontology will fall outside the computationally attractive OWL 2 fragments QL and RL. I suggest to avoid this and try to remain at least within OWL 2 RL such that sound and complete reasoning over rNews data is practically achievable even for large volumes of data.
A practical solution could be the introduction of a common super class to those classes that occur together in the disjunction and use this one in the domain and range definitions. For instance, this would lead to a super class of Comment and NewsItem.
Posted at 9:12 AM on May 31, 2011