I think using html and xhtml as examples of technology is where the first problem lies. As Charles said, vacuum tubes and transistors could be seen as very similar in functional terms but are very different in the technological means by which they achieve that fucntion, hence a technological detour makes sense as a concept - two totally different technological routes to the same functionality, one becomes a dead end and the other becomes the basis for future advance.
When you are operating in a virtual world, which is the appropriate analogy for the physics/chemistry side of the technological world?
Is it a case that function is king in computing and therefore functional equivalence is all that matters?
Or are rules & syntax the equivalent of physical/chemical mechanism? Or delivery method? Or...?
Alternatively again is it really about 'parentage' of concepts, i.e. transistors did not utilise the same scientific breakthroughs as vacuum tubes (at least the more recent ones, naturally if you go back far enough...) , so they are not closely related in technological terms? In that case I would say xhtml has xml as it mother and html as it father, so it is not really a detour?
As far as the 'view it in a browser and you can't see it' argument goes, I can't agree with that - very similar programming languages can't be compiled by each other's compilers, applications can't always open data files form previous/future versions, that is not sufficient to say the two are unconnected/unrelated?
Likewise 'same' or 'similar' are relative (and often also subjective) terms - no two dogs are the same, but two dogs of the same breed are often more similar than two dogs of different breeds, which depending on your criteria might be more similar than a given cat and dog, or then again may not...?
Last point, isn't history the only judge of a technological detour, in which case the question should be "will html prove to be a technological detour" - it isn't at the moment as there is no definitive shift from one to the other.