Thread
#data-architecture-discussion
    einat.orr

    einat.orr

    4 weeks ago
    Is that approximated reproducibility valuable in your opinion?
    h

    Hugh Nolan

    4 weeks ago
    Thank you @einat.orr for your thoughts! Yes, the approximate reproducibility is useful, we do scientific analysis so being able to reproduce historical results - exactly, or approximately - is very good. I suppose I could also write a separate "redactions" log which could be referenced as necessary.
    einat.orr

    einat.orr

    4 weeks ago
    The "redactions" log is a great idea. I love it :jumping-lakefs:
    You can find the commit-log here: https://docs.lakefs.io/reference/api.html under Refs, the third one 🙂