Ad-hominem:
http://www.sjsu.edu/depts/itl/graphics/adhom/adhom.html
The jump to the relationship "I knew a guy who was brilliant but..." can be construed as ad-hominem as how is it relevant to the topic of discussion except as an implied potential link to the mental state of the author (by association) so the work can then be dismissed rather than discussed.
The abstract is interesting, the journal is peer reviewed and prestigious:
http://www3.interscience.wiley.com/j...formation.html (no guarantee that everything is perfect but certainly helps limit nonsense).
I can not readily get to the underlying article so can not comment further.
Observations are observations, theory is theory, we try and match the two but of the two observations are what is kept if one is not lining up with the other. For example, just as Newtonian view of gravity is pretty good most of the time, so might be the case for relativity. Prior to Kepler there was a way of describing the orbits by many circles - still possible but not as simple or elegant as ellipses and then gravity to explain why ellipses. What helped all this were observations, e.g. Tycho & others.