According to Eric Dezenhall, a public relations advisor for the Professional and Scholarly Publishing Division of the Association of American Publishers, there is a potential problem with open access scientific publishing:
Specifically, according to Dezenhall's suggestions in a memo to the publishers that they should "develop simple messages (e.g., Public access equals government censorship; Scientific journals preserve the quality/pedigree of science; government seeking to nationalize science and be a publisher) for use by Coalition members."A government might possibly try censoring past results if the data were held in servers under the control of that government. (They might also try censoring present results but that goes along with the government funding in the first place.) For example, if a creationist or deep ecologist administration gets elected they might try removing some articles from the database.
We can prevent such censorship by having mirrors not under government control. They exist for at least one public-access organization, arXiv.org. People opposed to the possibility of censorship should recommend the multiplication of mirrors.
Multiple mirrors might be a little expensive now but, in accordance with Moore's Law, the cost should decline rapidly. Eventually, the mirrors might be supported by volunteer efforts similar to SETI@home.