研究问题：谁应为开放数据服务买单

结论：

The 2020 State of Open Data report provides an interesting lens to view how far open research has come, and to look at opportunities for improvement in data sharing. Every time we push science forward, we should also both reflect the past and predict the future benefits and challenges of our actions. 2020 has offered tests and trials like no other in my lifetime. This time has also shone a light into the gaps in our thinking as well as in our progress towards open research. This fifth edition of the State of Open Data report reveals the current thinking on open science from a global pool of over 4,500 respondents. We have the opportunity to simultaneously reflect on how this movement has transformed into practice as well as thoughts to consider on moving forward in this space. Over the past five years, the science ecosystem of researchers, librarians, publishers, institutions, funders, and others have embraced improving data sharing. Policies requiring more research transparency and data sharing have emerged alongside communities aiming to improve scientific scholarship. Yet, have the policies and open data conversations affected practice? Let us look at the effect of requiring data availability statements in publications – a mechanism to ostensibly make it easier to find data and encourage research data sharing. However, when one looks within a publication, there is frequently a statement such as ‘Data available upon request’, or to be either more polite or more cynical – ‘Data available upon reasonable request’. This may be why a majority of survey respondents want stronger mandates for data sharing, with enforcement. While data sharing policies have not fully translated into practice, the conversations and atmosphere of sharing research has changed. Covid-19 has illuminated the needs and capabilities in making science open and accessible and perhaps surprisingly, in doing so, suggested that science truly can be accelerated. The scientific community, now indoctrinated in open science, has embraced the principles of sharing their research openly. Disparate communities – academic, governmental, commercial – have cooperated to create an array of solutions needed to combat Covid-19. Researchers, repositories, funders, and more have independently and jointly made Covid-19 related scientific research freely and openly available with necessary protections for private data. The Research Data Alliance (RDA) convened over 200 global volunteers to report1 on the needs of data sharing culminating in a report released in June 2020 to guide data sharing during a pandemic. Open Science and data sharing practices cited in the RDA report range from adhering to the FAIR (Findable, Accessible, Interoperable and Reusable) principles to documenting methodologies, incentivising early publications and expediting review processes, implementing legal frameworks for cross-jurisdictional data sharing balanced with ethical and privacy considerations. These principles and recommendations are salient to open science and open data at any time. As we look to the future, what will be needed to intentionally improve open data? Three interrelated topics - trust, misuse, and equity - must continue to shape future conversations to enrich and protect our Open Data ecosystem. As discussed in the previous and current State of Open Data reports, the issue of trust and data misuse weigh significantly in the minds of many and in particular of researchers. Researchers have concerns in having data misused and of others finding errors in their data. Yet there are other types of trust and data misuse that have been in the spotlight this year. The algorithmic biases embedded in the architecture of this digital age continuously surface at an alarming rate. We know that the algorithms have been developed from biased data, but how much has this affected or will affect open data? Misuse of open data and open science practices can be intentional or unintentional, but its presence is undeniable. For example, GitHub was originally conceived for open code sharing and exchange but is now ‘misused’ to store scientific work one might argue belongs in repositories. Without formalities of metadata tagging and organisational structure, these open data are accessible but not necessarily findable or interoperable. Yet, the data are more available than sitting on a local computer. On a nefarious front, as politics and science have collided, misinformation efforts have infected the scientific structure as they have in politics. Questionable scientific ‘research’ has been uploaded on established research sharing platforms then highlighted in the news as a source of legitimate scientific information. There has also been a growing misuse of pre-published research when moving research into the public eye before the methods, results, and conclusions have been rigorously scrutinised. Thus, our natural science ecosystem has been misused outside the non-scientific community.

建议：

The 2020 State of Open Data report provides an interesting lens to view how far open research has come, and to look at opportunities for improvement in data sharing. Every time we push science forward, we should also both reflect the past and predict the future benefits and challenges of our actions. 2020 has offered tests and trials like no other in my lifetime. This time has also shone a light into the gaps in our thinking as well as in our progress towards open research. This fifth edition of the State of Open Data report reveals the current thinking on open science from a global pool of over 4,500 respondents. We have the opportunity to simultaneously reflect on how this movement has transformed into practice as well as thoughts to consider on moving forward in this space. Over the past five years, the science ecosystem of researchers, librarians, publishers, institutions, funders, and others have embraced improving data sharing. Policies requiring more research transparency and data sharing have emerged alongside communities aiming to improve scientific scholarship. Yet, have the policies and open data conversations affected practice? Let us look at the effect of requiring data availability statements in publications – a mechanism to ostensibly make it easier to find data and encourage research data sharing. However, when one looks within a publication, there is frequently a statement such as ‘Data available upon request’, or to be either more polite or more cynical – ‘Data available upon reasonable request’. This may be why a majority of survey respondents want stronger mandates for data sharing, with enforcement. While data sharing policies have not fully translated into practice, the conversations and atmosphere of sharing research has changed. Covid-19 has illuminated the needs and capabilities in making science open and accessible and perhaps surprisingly, in doing so, suggested that science truly can be accelerated. The scientific community, now indoctrinated in open science, has embraced the principles of sharing their research openly. Disparate communities – academic, governmental, commercial – have cooperated to create an array of solutions needed to combat Covid-19. Researchers, repositories, funders, and more have independently and jointly made Covid-19 related scientific research freely and openly available with necessary protections for private data. The Research Data Alliance (RDA) convened over 200 global volunteers to report1 on the needs of data sharing culminating in a report released in June 2020 to guide data sharing during a pandemic. Open Science and data sharing practices cited in the RDA report range from adhering to the FAIR (Findable, Accessible, Interoperable and Reusable) principles to documenting methodologies, incentivising early publications and expediting review processes, implementing legal frameworks for cross-jurisdictional data sharing balanced with ethical and privacy considerations. These principles and recommendations are salient to open science and open data at any time. As we look to the future, what will be needed to intentionally improve open data? Three interrelated topics - trust, misuse, and equity - must continue to shape future conversations to enrich and protect our Open Data ecosystem. As discussed in the previous and current State of Open Data reports, the issue of trust and data misuse weigh significantly in the minds of many and in particular of researchers. Researchers have concerns in having data misused and of others finding errors in their data. Yet there are other types of trust and data misuse that have been in the spotlight this year. The algorithmic biases embedded in the architecture of this digital age continuously surface at an alarming rate. We know that the algorithms have been developed from biased data, but how much has this affected or will affect open data? Misuse of open data and open science practices can be intentional or unintentional, but its presence is undeniable. For example, GitHub was originally conceived for open code sharing and exchange but is now ‘misused’ to store scientific work one might argue belongs in repositories. Without formalities of metadata tagging and organisational structure, these open data are accessible but not necessarily findable or interoperable. Yet, the data are more available than sitting on a local computer. On a nefarious front, as politics and science have collided, misinformation efforts have infected the scientific structure as they have in politics. Questionable scientific ‘research’ has been uploaded on established research sharing platforms then highlighted in the news as a source of legitimate scientific information. There has also been a growing misuse of pre-published research when moving research into the public eye before the methods, results, and conclusions have been rigorously scrutinised. Thus, our natural science ecosystem has been misused outside the non-scientific community.