ong ( on December 18, 2007, 00:17:18 ) If you distribute data, how do you make sure the same open access principles apply to people who reuse/redistribute it? http://sciencecommons.org/weblog/archives/2007/12/16/announcing-protocol-for-oa-data/

mgashler ( on June 22, 2009, 20:21:33 ) Someone once told me that copyright does not apply to factual data. As an example, my source told me that the look and layout of a particular phone book could be copyrighted, but the names and numbers in that phone book could be copied and redistributed freely. Was this faulty or dated information? If we specifically designate a bunch of datasets as "free", might we unintentionally set a precedent that actually throws away rights? Does human involvement make data non-factual? (ie, could I copyright the data in a phone book if a human was involved in collecting the data?) Does a collection of measurements count as non-factual because it is subject to the accuracy of the instruments that measured it? What if I consider the instrument noise to be part of the truth in which I am interested, thus making the data exactly correct? (This might be the case, for example, if I am trying to design a noise-tolerant learning algorithm.)

ong ( on June 24, 2009, 13:52:26 ) There is a fine line between what is factual and what is created. I do not know what is considered copyrightable, and I suspect that this definition changes every now and then, depending on which corporation wants something. The two recent "free data" licenses specifically throws away rights, putting data into the public domain. CC0 PDDL For academic research, I believe that this is a good way to go, since it would be natural for the data provider to "just request" that the users to cite some paper or other, and users would do so. Hence the data is put in the public domain, and the authors get maximum benefit. It is not clear what to do in industry or when one is collaborating with a partner who may want to make a profit from the data.

Post a Reply