The overall purpose of this position paper is to provide set of recommentions about:
THIS DOCUMENT IS AVAILABLE FOR PUBLIC REVIEW TILL THE 13th MARCH 2012.
Please add your comments to this Community Site item, or send them to email@example.com.
Thanks for an interesting paper. I wonder whether we should make a clearer distinction between data publishing and data sharing. When something is published, it cannot usually be pulled back. With sharing it can. One of the fundamentals of GBIF has been that data providers are in control and can in principle remove their data from the network, if they want. Now with publishing, is this anymore the case?
Compare this with the Morphbank system, where this notion of publishing has been taken literally. Once an image or other object has been published, it cannot be removed from the system anymore (well, except reinstalling the entire database from backup, which we have had to do a few times). This is to ensure that the URIs of Morphbank remain usable in real publications. The object can, however, be edited and annotated, if there is a problem.
To me it looks that this model of data publishing only applies for metadata. The dataset can still be changed or removed. The metadata of its existence perhaps cannot.
Hannu Saarenmaa 754 days ago
Since I recently spend some time evaluating licenses for data, I've read the "Intellectual Property" part. I agree with the recommendations: convention on citation + use (creative commons). At Canadensys, we have agreed to publish all our data under CC0. This should avoid any legal constraints for someone interested in using our data. I hope this encourages other organizations to use CC0 for their data.
Peter Desmet 725 days ago