Codata workshop

SESSION 2: SCIENTIFIC AND TECHNICAL DATA AND INFORMATION POLICY

• 
  Policy and legislation relating to archiving of publicly funded scientific data and information.
  
• 
  Promoting data sharing policies and procedures and institutional mechanisms for open access to scientific data and literature.
  
• 
  Developing convincing scenarios about the importance of science and technology, and related data requirements and issues, at the appropriate audience level
  

Chair: Dr Daisy Selematsela, National Research Foundation

Panellist: Dr Sospeter Muhongo (ICSU)

Recruitment

It is an indictment that so few African countries are CODATA members, as it should have been possible to recruit from among the participants at the 2005 workshop. It is necessary to be able to present a summary of the work of CODATA and the benefits of membership in order to recruit. The ICSU Regional Office for Africa, by comparison, has recruited ten African countries in its two-year existence.

With respect to the CODATA outreach programme, the strategy should not be for CODATA to attempt this alone, but to operate within existing systems. The point of entry should be African Union (AU) and NEPAD initiatives. The ICSU office can provide contact information and points of entry, particularly since Prof. Muhongo chairs the AU Commission on Science and Technology. Other systems that CODATA could consider using include the regional economic blocs, one of which is SADC. SADC recently took a decision to establish a science desk, which will probably be based with NEPAD within the Department of Science and Technology. The SADC bloc would also have to work with other blocs for North, West and East Africa. ICSU could put CODATA in contact with relevant organisation in all 53 African countries (or 54 if Western Sahara is admitted as an AU member).
However, the most effective outreach programme that CODATA could consider would be to work closely with the UN-proclaimed International Polar Year (IPY) programme for 2007/2008. The programme will be dealing with large quantities of data from a range of disciplines and has funds specifically for outreach. CODATA could consider submitting project proposals. South Africa has submitted three or four proposals. Information can be found on the website for the programme. South Africa has been given the mandate to chair the African initiative within the IPY Programme.
CODATA could consider submitting a proposal for funding to the UN-proclaimed International Year of Planet Earth, which will run from 2007–2009. This programme also has an outreach component.
Another possible partner is the Network of African Science Academies (NASAC), comprising the 18 African academies of science, including ASSAf. They might favourably consider a CODATA component.
The ICSU Regional Office for Africa is attracting the most powerful national ICSU member countries. The NRF subscribes to ICSU on behalf of South African scientists. The 29-strong scientific unions are even more powerful. All scientists around the globe are directly or indirectly affiliated to ICSU. ICSU is looking for partners to work with, and CODATA is welcome to approach the ICSU unions through the ICSU Regional Office for Africa.

Needs

From the African perspective, Africa needs validated sources of data to put at the disposal of policy makers and decision makers in order to challenge the validity of the findings of the World Bank and International Monetary Fund. Facts and figures are needed to support arguments and targets with respect to achieving the Millennium Development Goals (MDG). CODATA could take up the challenge of making such data available to counter data on Africa produced outside the African continent.

Other needs that CODATA could help meet include a database of African experts, including African experts in the Diaspora, possibly in conjunction with the Association of African Universities and the World Academy of Young Scientists (WAYS). WAYS has its head office in Budapest and is creating an office in the African region. WAYS Africa was launched a few months ago (with Dr Henry Roman of the CSIR as founding chairman). One of the advantages of knowing where African expatriates are based is that this assists African scientists to access well-functioning facilities for research. The Association of African Universities had a database up until the 1980s, but the data have not been updated and are now obsolete. Data are also needed for intellectual property purposes; for example, there are no comprehensive data on African rare flora and fauna species. Compiling such data is important in the light of the recently launched AU initiative in intellectual property.
If CODATA could link its initiatives to establish a database with the International Polar Year Programme and International Year of Planet Earth, it could consider presenting a research proposal to establish a database on African experts to the United Nations Educational, Scientific and Cultural Organisation (UNESCO) or the African Development Bank as potential partners.
The European Union (EU) has considerable funding available for African development, and ICT is a priority field.

Panellist: Robyn Glaser (Department of Science and Technology)

Organisation for Economic Cooperation and Development (OECD) principles and guidelines on access to research data

• 
  The OECD has produced Guidelines on Access to Research Data and the DST is researching their application in South Africa and exploring ways in which the DST and NRF could work together to facilitate increased data access in South Africa. The application of the guidelines will be further informed by workshops in August and September 2007. The DST is conducting a background study on international data policy, which will feed into the workshop. The workshop will hopefully result in an understanding of the application of the guidelines in South Africa as well as the legal, policy and compliance implications of creating a data-sharing infrastructure in the country.
  
• 
  Based on the example of the data-sharing model of the Human Genome project, it appears that an international data infrastructure would be an essential component of an international research agenda. This would become important for participation in large research programmes, such as those of the European Union framework programmes, in which South African researchers are already participating.
  
• 
  The DST is in the privileged position of observer at the OECD Committee on Scientific and Technological Policy and is the only African country to have this status, along with several other developing countries. This enables the DST to participate in a wide variety of policy issues that range from human capital development to specific science areas and the evaluation of science systems.
  
• 
  The DST recently developed a strategy for intellectual property rights (IPR) related to intellectual property (IP) generated by publicly funded research, which has gone to Cabinet.
  
• 
  The findings of enquiries by the DST among science councils and research institutions on the status of data access is that there little regulation of data access, which tends to happen informally, possibly at an institutional level according to the policies of the institution, or perhaps on the basis of contacts. There is very little transparency, and there is no formal mechanism for scientists in South Africa to deposit data.
  
• 
  The DST is working closely with the NRF, which is establishing the National Data Information and Curation Centre (NADICC) as the infrastructure for a potential data management system, backed up by policy support from the DST.
  
• 
  South Africa has much data that remains at the research institutions where it was generated but is not necessarily being mined, and the country is consequently incurring opportunity costs, which could even be regarded as a waste of public funding. The government wants optimum return on investment in public funding, because public funds are ultimately intended to have a socio-economic impact. By improving access and secondary access to data, it should be possible to reach that ultimate goal.
  

The purpose of the OECD guidelines is to:

• 
  Promote data access and sharing
  
• 
  Provide a commonly agreed upon framework of operational principles for the establishment of research data access arrangements in Member countries
  
• 
  Offer recommendations to member countries on how to improve the international research data sharing and distribution environment.
  

Research data are defined as factual records used as primary sources for scientific research, and that are necessary to validate research findings. A research data set constitutes a systematic, partial representation of the subject being investigated. This term does not cover the following: laboratory notebooks, preliminary analyses, and drafts of scientific papers, plans for future research, peer reviews, or personal communications with colleagues or physical objects (e.g. laboratory samples, strains of bacteria and test animals such as mice).

The OECD Guidelines on Access to Research Data propose a ‘data commons’ characterised by openness, flexibility and transparency. The practical applications include interoperability, formal institutional relationships, security and sustainability. The constraints include the protection of intellectual property rights, the right of first use of research data, the need to protect privacy and confidentiality, the right to restrict information concerning indigenous species and the protection of the rights of commercial exploitation.
The issues for consideration with respect to the guidelines include legal and policy issues, cultural and behavioural issues (education and reward structures), the technological infrastructure, interoperability and quality control, as well as dedicated budgetary planning and financial support.

Panellist: Paul Uhlir (US National Academies, CODATA USNC)

The presentation provided background on the institutional types of mechanisms for providing open access to various kinds of scientific information and tabled several proposals that arose from discussions at a similar workshop in Brazil for the Latin-American region at a CODATA workshop and an Inter-Academy Panel planning meeting the previous week. Two of the proposals in particular may be worth considering in the southern African region.

Emerging institutional paradigms for open and permanent access

CODATA launched the Global Information Commons for Science Initiatives at the World Summit on the Information Society.

What is an information commons?

• 
  Digital data and information originating principally from government or publicly funded sources
  
• 
  Made freely available for common use online
  
• 
  With the material in the public domain, or with only some rights reserved (using a common-use licenses, such as Creative Commons), or with full intellectual property rights, but under open access conditions
  
• 
  Typically organised thematically through an institutional mechanism.
  

Existing information commons models include:

• 
  Open-source software movement (e.g. Linux and numerous other programmes worldwide, many of which originated in academia for research applications)
  
• 
  Distributed grid computing or e-science (e.g. SETI@Home, LHC@Home)
  
• 
  Open data centres and archives (e.g. GenBank, space science data centres)
  
• 
  Federated open data networks (e.g. World Data Center system, NASA Distributed Active Archive Centers (DAACs), Global Biodiversity Information Facility, South African Environmental Observation Network);
  
• 
  Open access journals (e.g. Public Library of Science (PLoS), which provides access to more than 2500 scholarly journals, many of which are published in the developing world, Scientific Electronic Library Online (SciELO), Bioline International);
  
• 
  Open repositories for an institution's scholarly works (e.g. the Indian Institute for Science and many others globally)
  
• 
  Open repositories for publications in a specific subject area (e.g. the physics arXiv, CogPrints, PubMedCentral)
  
• 
  Free university curricula online (e.g. the MIT OpenCourseWare)
  
• 
  Discipline or applications commons, or open knowledge environments (e.g. conservation commons).
  

Advantages of information commons for science:

• 
  Facilitates transfer of information North–South and South–South
  
• 
  Promotes capacity building in developing countries
  
• 
  Promotes interdisciplinary, inter-sector, inter-institutional and international research and cooperation
  
• 
  Avoids duplication of research and promotes new research and new types of research
  
• 
  Reinforces open scientific inquiry and encourages diversity of analysis and opinion
  
• 
  Allows for the verification of previous results
  
• 
  Makes possible the testing of new or alternative hypotheses and methods of analysis
  
• 
  Facilitates the education of new researchers
  
• 
  Enables the exploration of topics not envisioned by the initial investigators
  
• 
  Facilitates automated digital knowledge discovery and diffusion
  
• 
  Generally helps to increase the research potential of digital technologies and information, thereby providing greater returns from the public investment in research
  
• 
  Many other advantages and justifications.
  

Barriers to creating information commons:

• 
  Implementation and acceptance of new policy and institutional frameworks
  
• 
  Development of adequate incentives at the individual, community, institutional and governmental levels
  
• 
  Long-term financial sustainability of different information commons models
  
• 
  Effective technical and organisational approaches
  
• 
  In all cases, the information commons must balance with legitimate countervailing values and legal restrictions (protection of national security, privacy, confidentiality and intellectual property rights).
  

Broad implications of excessive restrictions (economic, legal, technical) on access to and re-use of data and information from public sources:

• 
  Disadvantage and marginalisation of developing country or poor users (especially impacting poverty reduction efforts)
  
• 
  Significant lost opportunity costs, and the related failure to capture maximum value from public investment in public data collection activities, including geospatial data
  
• 
  Monopolisation problems exacerbated in database markets, both public and private
  
• 
  Higher transaction costs (not just cost of access)
  
• 
  Less effective international, inter-institutional and interdisciplinary cooperation using digital networks.
  

Openness thus should be the default rule, subject only to legitimate and well-justified exceptions.

Additional background reading (all available freely online):

• 
  Bits of Power: Issues in Global Access to Scientific Data (NAS, 1997)
  
• 
  The Role of S&T Data and Information in the Public Domain (NAS, 2003)
  
• 
  Reichman, J.H. and Paul F. Uhlir, ‘A Contractually Reconstructed Research Commons for Scientific Data in a Highly Protectionist Intellectual Property Environment, 66 Law & Contemporary Problems 315-462 (2003)
  
• 
  UNESCO Policy Guidelines for the Development and Promotion of Governmental Public Domain Information (2004)
  
• 
  Open Access and the Public Domain in Digital Data and Information for Science (NAS, 2004)
  
• 
  Strategies for Open Access to and Preservation of Scientific Data in China (NAS, 2006)
  

What are some potential realistic cooperative activities in Latin America to help address the challenges and barriers that have been identified, and that might be able to be adapted to the southern African context?

First proposed initiative

WHAT: There is lack of knowledge about the role and scope of intellectual property (IP), what types of works are protected and in what contexts, and when common-use licences are appropriate. The situation is complex and changing, even for professionals in government and research. National research funders do not have open access (OA) in their mission or mandate. How to license for common use is still largely unknown. IP policy and legislation are controlled by commercial interests, mostly by OECD multinationals. There is thus a need to promote partnerships for sound policy-making on IP and OA for public research.

HOW: Build capacity in IP and OA by creating a partnership with stakeholders among national level IP experts, ministry officials, research funding organisations, universities, libraries and OA publishers, together with international partners. Develop strategies and models to promote, communicate and implement OA for publicly funded research. Context-dependent rationales are the most effective approach, making the case from the self-interested perspectives of each community.
WHO: Such a partnership should include coalitions at the national level, as described above, and at the international level (e.g. CODATA could coordinate with Inter-Academy Panel (IAP), Creative Commons/Science Commons, Access to Knowledge (A2K), Open Content Alliance, OA publishers, library organisations, other NGOs and domain-specific programmes and organisations). Establish a working group to develop such a partnership, perhaps on an Open Content Alliance (OCA) model. Hold an organising meeting in Washington in 2007. Initiate a pilot project with Chile.

Second proposed initiative

WHAT: Ibero-Latin American Open Access Repositories (ILOAR). Establishing open archives for S&T literature and data in Latin America, building on successful models and open-source tools. Need to strengthen ‘human interoperability’ and convince all segments of the research community of open access to information.

HOW: Proposed open archives models based on Open Archives Initiative. Many open source software (OSS) programs are now available (e.g. Open Journal Systems (OJS), e-Prints, Open Conference System (OCS), DSpace, Fedora, etc.). Need to do research, distribution of tool kits, advocacy in both Portuguese and Spanish.
WHO: Implementation by Instituto Brasileiro de Informação em Ciência e Tecnologia (IBICT) and partners, with IAP/IANAS (Inter-Academy Panel/Inter-American Network of Academies of Science), UN GAID e-SDDC (United Nations Global Alliance for ICT and Development, Global Alliance for Enhancing Access to and Application of Scientific Data in Developing Countries, DRIVER (Digital Repository Infrastructure Vision for European Research) (in EU), RedCLARA.

Funding: UN GAID regional office, European Commission, IAP Programme, Moore Foundation supports adoption of Fedora, etc.

Panellist: Dr Xola Mati (ASSAf)

Overview of Scholarly Publishing Report and the implementation of its recommendations

• 
  The DST published a Strategic Approach to Research Publishing in South Africa. The report emanates from a broad process of ongoing consultation over a period of about four years with the stakeholder community concerned with knowledge input and output, notably the NRF, National Advisory Council on Innovation (NACI), Council on Higher Education (CHE) Higher Education Quality Committee (HEQC) as well as Higher Education South Africa (HESA) (which represents South Africa’s 21 universities), the Department of Education and the Department of Science and Technology.
  
• 
  The report has ten recommendations. ASSAf is mandated by its stakeholders to take the lead in coordinating the process of implementing the recommendations.
  
• 
  ASSAf has submitted a three-year implementation plan to DST (with the option for review and support in a subsequent period), which is under consideration, subject to conditions, including a memorandum of agreement (MoA) with identified partners.
  
• 
  There is an ongoing process of consultation with stakeholders with the aim of reaching multi-party agreement on a business model for the scholarly publishing system in South Africa and critiquing of the current model.
  
• 
  The intention is to launch an editors’ and publishers’ forum for the more than 200 accredited scholarly journals in South Africa (and in this regard, editors and publishers of accredited journals have been invited to a workshop planned for 25 July 2007).
  
• 
  ASSAf intends to propose a new accreditation model for better journals funded by the DST. Open access is one of the issues on which ASSAf will constantly engage with its stakeholders.
  
• 
  ASSAf has also launched a study on scholarly books at the request of the Department of Education (DoE). The DoE lacks policy for accrediting scholarly books.
  
• 
  ASSAf intends to engage with editors on open access versus a print model of scholarly publishing, based on the considerations of access and visibility, with the aim of reaching agreement on the direction for South Africa.
  
• 
  There are a number of institutional repositories in the country (including the universities of Pretoria, the Witwatersrand, Stellenbosch, KwaZulu-Natal and Rhodes), but not all institutions have the necessary resources and capacity to set up their own repositories. There is a need for a national repository, given the requirements of institutions without the resources, linked to NADICC. This would also serve to share data across institutions.
  
• 
  ASSAf produces the science magazine Quest as part of its initiative to promote science in education and as part of the recommendations of the report.
  
• 
  ASSAf wishes to achieve a measurable impact of the research sector and the necessary intervention to ensure quality scholarly publications in the country, in conjunction with its partners.
  

Panellist: Nontando Guwa (ASSAf)

National Virtual Information Centre (NAVIC) on scholarly publishing

One of the recommendations of the report (Strategic Approach to Research Publishing in South Africa) was to establish a national virtual information centre to hold information on all scholarly publications published by South African scholars (including both books and journals), whether published in South Africa or abroad, and to make these freely available online. ASSAf is not considering storing theses, dissertations or research data on NAVIC, although it realises that there are interfaces between them. The NRF will continue archiving theses and dissertations, while NADICC is looking into data curation. By serving on the NADICC steering committee, however, ASSAf is in a position to provide guidance to the NRF initiative on data curation. A gap analysis is being undertaken among stakeholders to ascertain the gaps that they would like to see the National Virtual Information Centre (NAVIC) filling, and lessons are being learnt by drawing on international expertise and local experience (e.g. SARIS and e-Research). A proposal will then be developed on the formation of NAVIC and how it will operate.

Panellist: Dr Wieland Gevers (ASSAf)

Principles of research publication and data sets

• 
  The report (Strategic Approach to Research Publishing in South Africa) stresses that the basic requirements of the scientific method (including publishing), built up over hundreds of years, have to be built into any new system. However technically sophisticated or pervasive in terms of access any new system may be, the system may become inoperable through trying to house too much information with insufficient organisation and doubtful reliability. The scholarly publishing system includes a number of constraints, including that (1) a researcher may not publish what has already been published, (2) the methods used must be presented in sufficient detail to allow others to repeat the work, (3) the work must have been reviewed by peers who are able to judge methodological problems, citation practice, logic and background, and (4) editorial discernment is necessary to assess the peer reviews. There are thus considerable constraints before new knowledge enters ‘the literature’.
  
• 
  The technical advances in research practice and publishing raise many issues with respect to databases, for example, in biochemistry papers, the authors will have been required by the editor to put most of their data into a URL and indicate in the paper where the data can be accessed. Some of these are organised and systematic (for instance, by GenBank), but others are random and possibly ephemeral. The work cannot be repeated if the data in an ephemeral data holding are no longer available after several years.
  
• 
  There are also issues related to the extent of the data in the public domain. The exclusions include notebooks, draft papers and peer review reports. The more pressing issue is what should be included in the public domain, following the important principles of first use and copyright.
  
• 
  The ASSAf initiatives are not simply designed to overcome the technical issues of open access, but strive to be grounded in academic good practice in order to address one of the primary problems, namely, that African science is generally ignored internationally as it is considered to be of poor quality. Even local researchers may first try to publish abroad and only resort to a local scholarly journal as a last resort.
  
• 
  South Africa needs strong journals that are well edited and publish internationally important papers. Young researchers need to be taught how to write papers. Unless these aspects are in place, open access will not help but will make South African researchers even more enslaved to simply considering the work done by others abroad.
  
• 
  ASSAf strives to embed its initiatives in efforts to strengthen science in Africa and address primary criticisms of African research as poor quality, which cause even the best results to be overlooked.
  

Panellist: Dr Andrew Kaniki (NRF)

• 
  The NRF has a project on national electronic theses and dissertations. In order to support the work done in publishing and the integrity of data, the NRF is also embarking on data rescue and data archiving initiatives, recognising that researchers are not data managers and generally do not have the time and skill to archive their data once the initial analysis has been completed.
  
• 
  The National Data Information Curation Centre is a virtual centre; it is not intended to become a central repository for data but to be distributed. The NRF chairs a joint project involving a number of partners – the University of Pretoria, CSIR (as observer), HSRC, NRF (bringing its experience of the South African Data Archive), Medical Research Council and others. A preliminary survey was conducted.
  
• 
  NADICC is already involved in the Data Interchange Standards Association (DISA) project, which entails the digitisation of data, but is also involved with journals.
  
• 
  The intention is to identify the gaps as well as the people who can teach metadata so that skills can be developed in standardising, managing and archiving data, while at the same time addressing issues of data use, data security, and systems for data sharing. There are ongoing initiatives in South Africa (for example, the national research and education network, NREN). Ultimately, science councils and research organisations will use the NREN for the transfer, exchange and sharing of data, although at present, standard Internet connectivity is used.
  
• 
  The NADICC initiative is thus to ensure proper data management and policies, bearing in mind what is available and tying in with publicly funded research and impacting on access. The issues include intellectual property, the fact that not all data need to be stored permanently, criteria with respect to whether a particular data set should be made openly accessible or controlled. The important thing is to identify the location of the original data for authenticity.
  
• 
  The project will also address the need to develop capacity in secondary data analysis.
  
• 
  The NRF, ASSAf and HSRC have invested funds for the initial stage of the project to identify the available skills and data resources and training requirements as the basis for connectivity and in the interests of more coordinated data archiving.
  

Comments on presentations

Comment: Researchers often have limited awareness even of the data available in their own organisations. The intention is not to create a central repository but a virtual system in which each member organisation takes responsibility for its own data.

Panellist: Prof. Steve Rossouw (CODATA SANC)

• 
  Given that the session is devoted to scientific and technical data and information policy, the issue of what CODATA is doing nationally and internationally in this regard arises. The IPY, for example, is a joint activity by ICSU and the World Meteorological Organisation, coordinated by a joint committee. There is a subcommittee on data policy, co-chaired by Mark Parsons of the World Data Center for Snow and Ice (Denver, Colorado) and Taco de Bruin (Netherlands). Prof. Steve Rossouw is a member of the subcommittee. During a meeting in Paris in January 2007, the question of training and capacity was discussed. Prof. Steve Rossouw commented that CODATA SANC would like to offer data training courses in the region. Both co-chairs were enthusiastic, although their main focus at the moment is data policy for IPY, as well as legacy data. IPY lasts until 2010, with most of the research completed in 2007/2008, but the focus thereafter will be to archive the data generated by the project and make it accessible. There is a worldwide need for training in data handling, curation and archiving to deal with the massive amounts of data being generated. Prof. Steve Rossouw undertook to raise the need for data courses in the region again at the next subcommittee meeting in Canada in September 2007.
  
• 
  CODATA is aware of the need for data policy and recently approved a task group for this purpose (of which Prof. Steve Rossouw is a member). The task group will also put pressure on the joint IPY subcommittee on data policy to begin offering courses in the region.
  
• 
  CODATA is thus considering various ways of advancing data policy and offering guidance both nationally and internationally.
  

Questions and discussion

• 
  Comment: There is support for an inventory of what research initiatives, researchers and data exist in SADC (especially in areas of interest to the whole SADC region, e.g. malaria).
  
• 
  Comment: In creating such an inventory, it is necessary to identify big data sets in various fields that are of value to SADC and make these known outside the narrow disciplinary group working in the field.
  
• 
  Comment: It is necessary to identify international data sets that are interest to SADC.
  
• 
  Comment: The same need to identify existing data resources was identified at the meeting in Latin America. This is thus a commonality of interest with Latin America. Once existing data holdings have been identified, strategies of access can be developed.
  
• 
  Response: Many institutions have already developed data repositories, and researchers in SADC may be surprised to discover how much is available in southern Africa.
  

• 
  Comment: It could not simply be assumed that there would be open access to data from all publicly funded research in South Africa, for people to use in any way they choose, as the data are subject to complex legislation and intellectual property rights. It is necessary to keep track of the data generated through publicly funded research, and for government to take possession of the data once the project has been completed in cases where the data lie unutilised by the institutions by which they were produced.
  
• 
  Response: It would be necessary to decide (1) whether SADA is a potential model for a regional superdata centre, or (2) whether the preferred approach would be to have a bottom-up repository of data curated by the holding institutions according to exacting and agreed standards, forming a network that could be harvested effectively in an umbrella system, in which any user would have open access to any data owner. There seems to be agreement that in the SADC context, the second model of network of interoperable, harvestable data sources would be preferable to the first.
  

• 
  Comment: With respect to the different models, a decade ago the objective was to create centralised data centres, but these are but costly in terms of funds, human capacity and risk. The new methodology is to create System of Systems for distributed data. Kofi Annan, the former UN secretary general, in talking to UN GAID, spoke of building a decentralised Network of Networks. This is very different from a national data centre.
  
• 
  Response: There are approximately 70 so-called ‘world data centres’ in as many countries. The data generated by the IPY programme covers the whole spectrum of data, including socio-economic data, and has wide application. The IPY data committee has to consider where the data can be deposited. An initial survey of the world data centres showed that about half of them are almost dormant. Centralised data centres for a specific discipline tend to become dormant where the necessary human and financial resources are not available. It is time to consider a system of systems approach that will make it possible to develop a new solution to storing data that are generated internationally.
  

• 
  Comment: There is a possibility in SADC of moving to real open access. Given the obstacles with respect to intellectual property, sensitive data and expenses of accessing international data sources, there is a need for agreements on data access.
  

• 
  Comment: In the area of scholarly publishing, there are well-understood procedures of good practice; even if these are not always observed, they are aspired to. Are there similar quality assurance mechanisms for data archives?
  
• 
  Response: There are data curation strategies and best practices in different disciplines (e.g. baseline geospatial data sets, which are typically held by government agencies, but are largely inaccessible in many countries, which is a barrier to development as they comprise foundational data). There are peer-reviewed data sets with associated methodologies that could be more broadly promoted. This relates to the notion of data management training and academies.
  

• 
  Comment: Data users should not be overlooked. There is a need to train users from different fields to use data centres, starting with policy makers and funders.
  
• 
  Comment: In the definition of publicly funded data, data produced by UN agencies, intergovernmental sources and NGOs that fund public good activities should be included.
  
• 
  Comment: Statistics SA has set up a hierarchy of statistics based on the degree of reliability, and other government departments could consider a similar approach.
  
• 
  Comment: Much foundational geospatial data does not result from scientific research but is used by scientists (e.g. satellite imagery and mapping). Much satellite data, for example, is free to research institutions.
  
• 
  Comment: The development of capacity for secondary data analysis should be emphasised (e.g. reworked census data). The research community cannot continue to rely on external funding for data systems. Buy-in from government would be required for sustainability to maintain a system of systems. It is also necessary to make data user-friendly for government decision-making (e.g. livestock census).
  

• 
  Comment: Do SADC countries have legislation governing free access to research data by individuals?
  
• 
  Response: South Africa has a dual research funding system. There is input-driven research funded via the NRF, and there is output-driven research funded by grants from the Department of Education on the basis of publication. The scholarly journals make it difficult for the output of publicly funded research to be publicly accessible. The users either have to buy or subscribe to a journal, the publishers of which become the owner of the product, which was publicly funded. The system should be challenged in order to ensure that publicly funded research is freely available.
  
• 
  Response: South Africa has the Promotion of Access to Information Act (Act No. 2 of 2000). However, the Act deals only with public information and has many exclusions. The application of the equivalent Act in the United Kingdom is breaking down because of the numerous exclusions.
  
• 
  Response: In the USA, there is the Freedom of Information Act, with the Paperwork Reduction Act and regulations of the White House Office of Management and Budget, such as Circular A-130 that promotes access to government information. At the more specific and local levels, many agencies and universities have their own policies and approaches relating specifically to research data and information. This is a very decentralised approach.
  
• 
  Response: China started a data-sharing programme five years ago. The panel to bring stakeholders and experts together to draft legislation has not yet been successful. Issues of national security are a concern. Government does not want to be fully transparent. The legislation remains unrealised, but the Regulation of Access to Government Information is to be published in China.
  

• 
  Comment: The task group should be more focused on moving forward, as it is almost two years since the 2005 workshop at which the recommendations were made. It is necessary to decide which data to deal with. Access to research publications is no longer a major problem. In the US and several EU countries, policy is being developed that publications based on publicly funded research should be made openly available, at least within a specified period of 6–12 months. The profile of the journal and the individual scientist is built by citations, including self-citations, which can become something of a game. Unpublished data may be of both good and poor quality, but remain untested beyond the laboratory in which they were generated. The task group will have to decide whether it wishes to deal with unpublished data or not.
  
• 
  Comment: The task group should thus define the way in which it wishes to proceed and then prioritise its initiatives. For SADC, health may be a natural priority, given that it was a focus of the 2005 workshop, as well as for strategic reasons and access to funding, since five of the Millennium Development Goals involve health. Health is very broad and complex, and within this area, further prioritise is possible (e.g. malaria, IP related to medicinal plants). The system of systems approach would be best, taking into account the legal frameworks, the IP regime in each country, and the differences in Africa between the French-speaking and English-speaking IP blocs as well as the recently established Pan-African IP regime and the ongoing debate on whether it replaces existing ones. It would thus be possible for the task group to start on a small scale and gradually escalate its efforts.
  
• 
  Comment: With a view to the ICSU general assembly in Maputo, Mozambique in October 2008, strategically CODATA should motivate inclusion in the programme to demonstrate its relevance to Africa. As part of the programme, scientists will be taken to the different provinces in Mozambique to interact with the scientists on the ground. CODATA could consider sending a representative to each province to inform scientists how they can interact with data. A day will be devoted to Science in Africa. CODATA needs to agree on what to showcase and what products it could demonstrate by then.
  
• 
  Comment: Ecosystems, water, the environment and natural disasters are other possible areas of focus. The task group should identify data holders, data users and their needs and form networks. The Latin American experience can show the way. They identified biodiversity, sustainable energy, water, IP, science and mathematics education and capacity building as priority needs. UN GAID identified disaster mitigation, poverty reduction and public health as priorities. Areas of common interest could be identified as focal points for the development of a data centre.
  
• 
  Comment: An important dimension from Latin America was the emphasis on interdisciplinary research, as well as networking and collaboration between social and natural scientists.