the future of documenting science: research papers & datasets
Posted: Fri Jun 26, 2026 6:31 am
This was raised here (viewtopic.php?t=8) but I think it makes sense to give it its own thread under the meta science label. Would love to hear any and all thoughts about documenting and publishing science.
Some things I think could be improved:
[1] science builds on top of science: we need a way to trace the lineage of research. For example, if something is falsified, it would be a useful way to flag the papers that depend on that work. (citations / references do that to some extent, but they are not designed for this purpose.)
[2] flexibility e.g. adding new results / findings that either replicate, contradict or are inconclusive with respect to old ones (this includes a dynamic data model to accommodate adding new variables.) when Afiq and I spoke about this he had a good analogy for this; twitter threads.
[3] standards!!!! something like what GAAP (Generally Accepted Accounting Principles) is to financial accounting. I think part of what makes standards in accounting work is that it uses the same language of accounting both internally (managerial accounting) and externally (financial accounting), where internal accounts have more flexibility and can accommodate nuances, but they still both use the same foundation. the science version of this would work in the same way.
Some things I think could be improved:
[1] science builds on top of science: we need a way to trace the lineage of research. For example, if something is falsified, it would be a useful way to flag the papers that depend on that work. (citations / references do that to some extent, but they are not designed for this purpose.)
[2] flexibility e.g. adding new results / findings that either replicate, contradict or are inconclusive with respect to old ones (this includes a dynamic data model to accommodate adding new variables.) when Afiq and I spoke about this he had a good analogy for this; twitter threads.
[3] standards!!!! something like what GAAP (Generally Accepted Accounting Principles) is to financial accounting. I think part of what makes standards in accounting work is that it uses the same language of accounting both internally (managerial accounting) and externally (financial accounting), where internal accounts have more flexibility and can accommodate nuances, but they still both use the same foundation. the science version of this would work in the same way.