I agree with the principle in theory. Of course outright banning anything "offensive" without clearly defining what is and is not offensive is very open to interpretation and can be abused. But when it comes to writing rules regarding offensive communication and harassment I think it's better to be over-zealous and treat all incidents on a case by case basis, especially in organisations where they're more concerned with incidents being made public and the organisation getting bad PR.
Also an important point is the "cannot be considered offensive" rule is an IT policy, not a general university one. Every employer I've had has made me sign an acceptable use policy from IT that says (among other things) "don't send offensive emails". It's not that much different. As I said in the original comment, the site is an oversimplification of a complex problem, and generic IT policies shouldn't really be considered anti freedom of speech for various reasons.