zednis
commented
8 years ago I would say that dataset's are not required to have an author. In some cases it may make sense for a dataset creator to have the role "author", but not in all cases and not in the example you have stated.
How about, creators and contributors can have additional role information (I believe we are currently going down this path) and that "author" is a potential role for a creator, but not required.
mrpatrickwest
We've said no in previous meetings. Marshall and Hao created the datasets from the PDF documents for the data rescue boundary activity. In addition to being the creators of the dataset they are the prov agents who ran the software to grab the data tables from the PDF in order to generate the dataset.
According to Elizabeth Ferriss, who works with Kerstin Lehnert, they would be the digitizers, but not the authors of the dataset.
Could we infer that the datasets derived from publications are authored by the authors of the publication. Probably not because the authors of the publication may have gotten the table from a dataset that they did not create or contribute to.
Just like we can't infer that the authors of publications that reference or cite datasets are the authors fo those datasets.
So can we definitively say here that we will not represent authors of a dataset?