adswa
commented
3 years ago hey @mih, here is my approach for the debconf talk. My general idea is to center it around the needs of a researcher.
I start with an overview of cool features of my field:
- a culture of open data,
- large open source software ecosystem,
- many services for hosting or storing or sharing code/data/,
- community efforts towards standards,
- interesting patient data)
and the difficulties that arise nevertheless:
- a lot of data but few RDM skills,
- great software and analysis tools but intransparent or complex analyses and unstable results across software versions,
- Few generic or interoperable workflows/service usage across institutes much rather isolated solutions that hinder collaboration,
- the need to update to latest standards once it evolves,
- data privacy issues around patient data prevents data sharing.
I then introduce DataLad and center on four features (version control, data transport, interoperability, provenance capture) and highlight how these help with open science. I'll try to make a point that the original idea behind datalad was to enable a similar level of tooling and culture for the distribution and version control of data as it is present for open source software development (not too prominent in the slides (yet), but the angle I have in my notes).
Let me know if the overall approach is fine with you. I'll also ask the others in next week's call whether they are fine with it. Current slides are in this PR.
This closes #32 and is a draft on a DebConf talk entitled "DataLad - Decentralized Management of Digital Objects for Open Science"
Todo: