Following the release of the 1000 Functional Connectomes Project manuscript addressing the feasibility of R-fMRI data-sharing and data-mining, an opinion piece describing a similar effort from the task-based fMRI literature almost a decade ago, was brought to our attention. Titled “Databasing fMRI studies - towards a 'discovery science' of brain function” by JD Van Horn and MS Gazzaniga in Nature Reviews Neuroscience (2002; 3(4): 314-8), the piece does an excellent job of summarizing the challenges that arise in pursuing high-throughput imaging (the pdf is accessible at http://www.nature.com/nrn/journal/v3/n4/...). We thought it worth drawing attention to this pioneering attempt to shift the task-based imaging community toward discovery science.