[odc-discuss] qualities of research data in open licensing

Jo Walsh jo.walsh at ed.ac.uk
Sat Jul 17 23:25:36 UTC 2010

Regarding this question of whether public domain style or share-alike  
style licences are more appropriate for research data:

I would say that in the large part, at the core, the sciences and arts  
and humanities work with different qualities of data - broadly, data  
that is independent of time and data that is not.

Open science has dominated the discussion of open data for research as  
yet - because very early and obvious impulses for open data came from  
the sciences, along with a well-developed tradition of activism.

To generalise yet more wildly, science deals with data that is. Once a  
set of observations is complete, from sensors, measurements during an  
experiment, or enumeration of bits of a genome - the data is done;  
there's no idea of going back and correcting it.

In "arts and humanities" data more often is subject to revision,  
change in nature as well as augmentation. New imaging techniques  
reveal that ancient tablets have been misread; concordance of more  
texts reveal placenames and jargon being commonly used in different  
ways at different times than previously thought; pottery shards have  
different cultural meanings in the light of later discoveries of  
similar structures.

The point: a ShareAlike license may be more valuable in the humanities  
just because of the quality of the data concerned; ethical and  
political considerations about data sharing aside. A corrected version  
of a data source may easily become a more canonical source; for the  
originators not only to have full credit but to benefit from others  
work in contributions back, seems critical to development across  
disciplines, indeed to the worthwhileness at all of future work.

Corrections, clarifications welcome.


Jo Walsh

Unlock places: http://unlock.edina.ac.uk/
EDINA National Datacentre
tel: +44 (0)131 650 2973

The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.

More information about the odc-discuss mailing list