[jmalicki]

It's solvable if publishing the dataset counts as a paper, and citations of the dataset which should be required count as citations for e.g. tenure.

For example, ImageNet for machine learning is a very expensive and difficult data set to produce that has resulted in revolutionary advances in machine learning. And people build models on it, cite their results as evidence their models are good, and cite the paper.

[nicoco]

This is an interesting idea. Although I am afraid that publishing a dataset, even a good one, will not be considered "real science" by our (broken) institutions.