Consider the source
Hi again, notknowing,
I strongly advise you to be VERY wary of what you read in New Scientist! This was once a fine magazine, but in recent years has become increasingly infested with uncritical and ill-informed articles which have terribly misled their readers. You might recall the recent flap over Justin Mullin's article in NS on claims by one Roger Shawyer, which has been widely (and justifiably) criticized by many physicists. Indeed, I just noticed a new post by Marc Millis in sci.physics.research, in which he describes why he dismissed Shawyer's claims without a second glance. (Millis once headed the now defunct Breakthrough Propulsion Physics Program at NASA, so he certainly cannot be accused of being unwilling to consider even extremely outre ideas!) Other physicists have offered detailed rebuttals, however--- probably more detailed than the subject really deserves, in fact.
I think you misunderstood coalquay404's comment. His point (and I have noticed the same phenomenon he did) is, I think, that arXiv eprints which are not formatted in Tex (or LaTeX) are much more likely to be of poor quality. No-one claimed this is a law of Nature; rather, we are talking about statistical likelihoods. Some archive eprints formated in MS word are valuable (if rather clunky in appearance); many formatted using TeX are Dreck. He offered a rough rule of thumb (and I second the advice); take it for what it is worth.
As for the Tajmar et al. eprints you wanted to discuss, any claims that gtr is off by 17 orders of magnitude should always be treated with great caution. Extraordinary claims require extraordinary evidence, and as Tajmar et al. themselves acknowledge, their work has not yet been confirmed by other groups. In addition to the lessons of history, there are also some specific reasons to be suspicious of their claims, but I don't wish to get into that: I just wanted to add my voice to those urging general caution concerning claims to have found a grossly incorrect laboratory scale violation of gtr.
I wish it were not necessary for me to add that I am NOT claiming "gtr is true" (indeed, I hold that such claims don't even make sense). I am claiming that gtr might turn out to give incorrect predictions concerning certain phenomena (indeed, past history and specific well established physical reasoning both imply that we should EXPECT this to happen at some point), but Tajmar et al. are not likely (in my estimation) to have produced the first clear experimental limitation on the validity of gtr.
Chris Hillman