Abstract
The contribution of the data paper publishing paradigm to the knowledge generation and validation processes is becoming substantial and pivotal. In this paper, through the information-processing perspective of Mindsponge Theory, we discuss how the data article publishing system serves as a filtering mechanism for quality control of the increasingly chaotic datasphere. The overemphasis on machine-actionality and technical standards presents some shortcomings and limitations of the data article publishing system, such as the lack of consideration of humanistic values, radical race for big data, and inadequate use of expertise in data evaluation. Without addressing the shortcomings and limitations, the reusability of data will be hindered, and scientific investment to facilitate data sharing will be wasted. Thus, we suggest that the current data paper publishing paradigm needs to be updated with a new philosophy of data.