Metadata generation for animals, experiments, equipment, and reagents.
Rainer Rilke's Book of Hours has a section that I think can beautifully describe how ideas are encapsulated in data.
...
There I can find, as in old letters,
the days of my life, already lived,
and held like a legend, and understood.
Then the knowing comes: I can open
to another life that's wide and timeless
So I am sometimes like a tree
rustling over a gravesite
and making real the dream
of the one its living roots
embrace:
a dream once lost
among sorrows and songs
-- Rainer Rilke's Book of Hours (Translated by Anita Barrows and Joanna Macy)
At the moment, many studies likely leave those who follow in our footsteps feeling quite lonely and lost. Once data has been compiled and its fruits distributed through publication, the information therein acts as new fertilizer upon which future ideas may grow and bear fruit. Datasets are curated through immense effort and care. The tools used or developed often take years to master and understand. Unfortunately, much of scientific infrastructure does not adequtely support or integrate frameworks that enable sharing of information freely and easily. If we all developed tools and metadata together as a community, we could offer one another letters that describe our data as Rilke describes his thoughts. Simpler to understand and easier to share information without the necessity to parse the difficult prose contained in articles. From understanding comes knowing. If the temporary gravesite of an idea, experiment, or dream were easier to investigate and grow from, we might all reclaim lost dreams of discovery.
This repository is intended to follow the words of Jason Scott.
The goal is to leave ln2fs, or "Love Notes to the Future", for both future members of the lab and those who want to understand data we generate easier. Integrating tools like DataJoint and Neurodata Without Borders will be much easier if there was a lab-wide way of programmatically accessing lightweight metadata about subjects, equipment, reagents, and experiments. Ideally, this repo could help make such metadata easy to generate in a structured manner that's flexible across many projects and experiment types.
This is heavily inspired by Jonny Saunders' mentorship, their work through Autopilot, and their advocacy for helping make (neuro)science accessible and shareable for everyone. See here for an awesome piece of writing on the subject.
This is also inspired by a conversation I had with @anoushehbsuroosh in the Tye Lab. I will not perform a rap song about this unfortunately.