Klint Finley at ReadWriteWeb discusses a report from the Ontario Information and Privacy Commissioner that asks if anonymization preserves privacy against attempts to re-identification. The report concludes that anonymization remains an important safeguard, and that the risks of re-identification of individuals within a crowd of anonymized data are lower than many fear.
That's good news. All guarantees of anonymity are probabilistic, so having some evidence that if done right, anonymization can preserve privacy to a reasonable degree is helpful as libraries try to gain some benefits - primarily, serving their users better - from anonymized data. It'll be even better as best practices, backed by math and experience, continue to emerge.