luis_in_brief, 1 month ago

@ldodds the paper is directionally correct; I wrote about some of the same issues here: https://www.openml.fyi/open-data-entities/

ldodds, 1 month ago

@luis_in_brief thanks. I skimmed the abstract but it didn't immediately look to be recommending "entities"/"institutions", but will give it a proper read.

luis_in_brief, 1 month ago

@ldodds it gets there at the end, sort of. Not exactly the most to-the-point paper though ;)

The critical thing we’re both agreed on is that unilateral public licensing isn’t going to cut it, and we’re seeing experimentation along those lines with things like this: https://takes.jamesomalley.co.uk/p/how-to-use-nhs-data-for-scientific

ldodds, 1 month ago

@luis_in_brief also, I was pointed at this group recently, who were new to me.

https://tial.org

Institutional design seems key and I'm not sure I've seen a lot of deep experience in that area in open data efforts that I've seen so far.

luis_in_brief, 1 month ago

@ldodds hadn’t seen TIAL, very intriguing. Hard for me to speak directly to open data, but open source definitely isn’t great at that and open data (from a distance) seems to be very set on reinventing all our wheels without learning from any of our mistakes 😬