On a more serious note, perhaps when it starts to become clear that the process of writing the paper will help to clarify the scientific message for the authors themselves?

Once you have a good new acronym?

Congrats, Eli!

Maybe an opportunity to coin a term? How about "number maximalism"?

Brilliant!

How does one find deconflicted reviewers in this case?

🤔

I understand that these considerations would be difficult to formalize, since it is probably easy to use these as reasons to simply say no to everyone. But, ultimately we have to face that devices cannot be cloned, unlike code.

Beyond that, we would want some confidence that providing access leads to something meaningful, not just distribution for distribution's sake. Access should be collaborative in nature: intentions and plans understood and agreed in advance, and some involvement throughout.

A lot to discuss here, so just a couple thoughts off the top of my head. I would prefer to wait until any manuscripts being written about data on those devices are published, in case something comes up that requires remeasuring or new measurements.

Maybe. Who would pay for and manage it? Lots of logistical and practical issues to consider there that don't exist with code and data repositories.

For that reason, I do think sharing of physical samples is best done collaboratively, which creates other burdens. Happy to discuss this more!

I tend to think other matters present higher barriers, and competitive advantage is more in skills. Typically, code can be reused and extended in ways and with levels of reliability that physical samples can't (experimental no-cloning theorem?).

What is the argument against sharing devices?