I've read Michael Jackson's article and I didn't interpret it as saying that people using the term isomorphic are outright wrong. Some people are making this assertion and that is what I am trying to refute here.
I disagree with Michael's statement that two things that are the same cannot be isomorphic. The very definition he cites does not exclude the possibly of those things being the same!
"Corresponding or in similar form or relations"
But even then the assertion that the server and client side are exactly identical is not correct. Typically such code bases will have slightly different entry points for server and client. Additionally, React 0.14 beta is now splitting the DOM and String renders.
For some 'Universal' is attractive, because they want to use Javascript everywhere. I get that. If you like that, please go ahead and call it 'Universal'.
However, what matters equally is having an abstraction which allows pure functions to declaratively describe the form of components based on immutable data, no matter what the computational model of the target. It is arguably the abstraction that is the isomorphism and code that uses it could be called isomorphic. That could be stretching the term as per the above definition, but that doesn't mean that people who use it are wrong.
I think React provides the necessary abstraction. I have total respect for what the React core team and the community have achieved. I understand that Michael Jackson and Ryan Florence are motivated by wanting to get more people using React. This is a very good thing.
However, other people may be more interested in promoting the ideas behind React rather than React or Javascript specifically. It is these people I am trying to defend.