I have an idea for my research workflow for some time now: https://github.com/replikativ/datahike/issues/97
I would be interested what your opinions are about it?
I would use Datahike inside of my literate programming setup (org-mode) to conduct scientific research on large scale distributed machine learning infrastructure.
@csm301’s GC work will come in handy for that.
IMO, that sounds like a killer feature for datahike. I could imagine forking a database for different use cases, and having independent transactions take place, then possibly have a merge back together to get a single source of truth. It sounds like a good way to handle the limitation of having a single transactor.
Found an interesting article on hacker news recently about local-first software: https://news.ycombinator.com/item?id=21581444
Is this kind of software possible to build with datahike today?