This is the abstract I have submiitted for the Berlin-5 meeting : “Berlin 5 Open Access: From Practice to Impact: Consequences of Knowledge Dissemination”
Open Access to Research Data: surmountable challenges
Many scientists and organisations have recognised the power and importance of “Data-driven Science” where existing data is a primary resource in scientific research. In some communities (astronomy, particle physics, and some biosciences) this type of work flourishes and the primary challenges are technical – size, complexity, metadata, automation, etc. In many fields however, and almost all multidisciplinary endeavours the major obstacle is finding scattered, heterogeneous data. Many of the data first occur in scholarly publications and, while they can be interpreted and understood in low volume by humans, are poorly presented for re-use by machines. As an example, over 1 million new chemical compounds are published yearly, but are scattered through hundreds or thousands of journals.
In principle this could be solved by robotic indexing and the use of search engines. In chemistry, for example, we have developed text-mining techniques which can recognise as chemicals over 80% of terms in mainstream publications, and identify a similar percentage. Our tools could rapidly index the scientific chemical web and add significant semantic value.
The biggest problem, however, is that many publishers forbid or obstruct this activity. Most chemistry journals are closed and thereby immediately inaccessible to many. Even for subscribers there are usually lengthy licences which are fuzzy and difficult even for experts to interpret. There is an imbedded fear of offending publishers’ conditions either because of breaking copyright (even unintentionally) or being cut off by the publishers machinery (anecdotally very common). Many publishers specifically forbid robotic indexing.
The problem is solved for any “Open Access” publisher that adopts the spirit of the BBB declarations. Taken logically BBB requires that all content can be indexed and downloaded without permission. Unfortunately many publishers use “Open Access” but decorate their web site with additional licence conditions which are logically and ethically incompatible.
The label “Open Access” is a weak tool when describing access to, and re-use of, data. I and others have promoted the term “Open Data” (http://en.wikipedia.org/wiki/Open_data and references therein) to describe the need to consider data as a critical resource which needs political and legal activity. The use of Creative/Science Commons licences is extremely valuable but will need refinement as the principles of Open Access and Open Source do not translate automatically to data.
I shall give demonstrations of Open Data resources and outline some of the issues that the scholarly community must address rapidly if we are not to be impoverished by the “land grabbers” in the digital dataverse. We need a radical rethink of conventional information protection and need to be braver and more outspoken.
[Note: This was written pre-PRISM. I am concerned that if PRISM has any traction it will impact on Open Data as well as Open Access and will blog this later.]