Several blogs have discussed the question whether the crackpots, fringe scientists and especially "professional science critics" should have a free access to scientific resources such as arXiv.org. The boundaries between these groups on one side and scientists on the other side may sometimes be fuzzy, but when you see one, you usually know it's a crackpot.
Greetings to Chris Oakley, Danny Ross Lunsford a.k.a. Iman Zumbal, Quantoken, Tony Smith, secret milkshake, MathPhys, Pudding, and many others.
My answer is a pretty resounding No. We live in a world where the free market of ideas can show its invisible hand. Virtually anything can be written by the bloggers and anything written by the bloggers may be read by others (except for those who live behind the Great Firewall of China). Many different ideas are proposed on the streets, at homes, at schools, and at many other places. Ideas are born at different levels and their strength is evaluated by other people who think about them and test them.
But the purpose of the scientific journals and their electronic counterparts is very different from the desire to be the broadest possible arena for all people in the world to express their opinions about science. The purpose of scientific journals is to present scientific results. In many cases, journals only want to present the results of the highest available quality. The degree of "elitism" in other scientific resources may be less stringent but it should never decrease to zero because treating all ideas, right or wrong, to be on equal footing is in contradiction with the defining properties of science.
The crackpots who keep on saying that Einstein had to be wrong and theoretical physicists who continue to use Einstein's insights in their work are subjects of a huge conspiracy have no business at arXiv.org. And a similar comment applies not only to special relativity from 1905. If a science-hater keeps on writing similarly vacuous critiques of any other field in science without having any technical arguments, such a person simply should not be able to publish these things in scientific journals or their electronic counterparts. William Dembski should not be allowed to flood the archives of biological preprints with links to his silly creationist diatribes, and I could give you many other examples.
Obviously, we primarily talk about a very specific example. Everyone who understands high-energy physics at the level of an advanced graduate student or better knows very well that the blog "Not Even Wrong" has virtually no physical content whatsoever. The statements on "Not Even Wrong" are usually political, scientifically vacuous statements, and almost every attempt to write something that is slightly technical results in a nonsense because the author has not simply mastered modern high-energy physics, not even at the introductory level.
Obviously, many people outside the field are often confused about completely elementary questions and they cannot distinguish the content of "Not Even Wrong" from science. But these people simply shouldn't determine policies of arXiv.org because they are laymen or outsiders. The server arXiv.org was created as a tool to exchange scientific results between scientists who work in particular fields of science. "Not Even Wrong" has no scientific results. The only reason why the owner of the blog wants his trackbacks to appear at arXiv.org is to increase the number of visits of that blog and to give it a feeling of "scientific legitimacy".
That would be a fraud, however, and the number of visits to similar websites should certainly not become a goal of arXiv.org. From a scientific viewpoint, "Not Even Wrong" is trash - and this statement is not just a random opinion but an assertion that can be objectively confirmed. It is one of the defining features of science that the results that have been shown wrong, defective, vacuous, internally inconsistent, and mathematically flawed as well as those that have led to no new insights for decades are eliminated, which is why it's the job of the arXiv.org to try to protect the scientific server from an uncontrollable inflow of links to scientifically defective resources.
Another opinion I disagree with is that the scientific legitimacy or even correctness of scientific results is determined by a university affiliation. There has always been an efficient approximate filter working for arXiv.org that prefers senders with academic affiliations, as determined by their e-mail addresses. But it is certainly not a universal rule that people at universities have a scientific approach to questions while the people outside must be crackpots. I know very many counterexamples on both sides although, truth to be said, there is certainly some positive correlation between university affiliation and expertise in a field of science.
It's great to have all possible ideas and opinions around and accessible, but there should still be a difference between a scientific journal or an electronic archive of preprints on one side, and a web server designed to entertain and please crackpots on the other side.