It is often said that science is self-correcting, but it is usually more accurate to add “in the long run” to that statement.
Take, for example, this retraction of a 10-year-old paper in Entropy that had been questioned since 2005. Here’s the notice for “Statistical Convergent Topological Sequence Entropy Maps of the Circle:”
The editors were made aware that a paper published in Entropy in 2004  may have plagiarized an earlier paper by Roman Hric published in 2000 . After checking with specialized plagiarism software, we found that this claim is indeed correct and almost the entire paper is a verbatim copy of the earlier one. After confirmation of this fact, the editors of Entropy have decided to retract the paper immediately.
We would like to apologize to the readers of the journal that it took so many years to notice this error and to retract the paper. Apparently there is a comment on MathSciNet (http://www.ams.org/mathscinet/) since 2005 that points out this case of possible plagiarism , however the editorial office was not aware of this until recently. We request readers of the journal to directly get in touch with the editorial office and the editors of the journal for similar cases in the future, so that they can be handled promptly.
1. Aydin, B. Statistical Convergent Topological Sequence Entropy Maps of the Circle. Entropy 2004, 6, 257–261.
2. Hric, R. Topological sequence entropy for maps of the circle. Comment. Math. Univ. Carolin. 2000, 41, 53–59.
3. MathSciNet, MR2082710 (2005f:37075), http://www.ams.org/mathscinet getitem?mr=2082710
The paper has yet to be cited, according to Thomson Scientific’s Web of Knowledge.
Hat tip: William T.A. Harrison