2019-10-14: Development Meeting

Kacper, Tommy, Mike L, Craig, Elias, Mike H, Tim

Agenda

  • Updates
  • Discussion
    • XRAS proposal - [name=Kacper] [name=Craig]
    • Reproducible Run - [name=Craig]
      • Single tale, multiple runs
        • Each reproducible run is basically a tag that you can revert to
          • State of your workspace and metadata
        • Bertram: In a given workspace you can have the output of multiple run
          • Assumed this is the way it would be
          • No one else thought this way
        • Comparability
      • Tim: we need a fundamental model of what a Tale is
        • ER diagram
      • Distinguish between version and RR
        • Kacper: Any published RR is a version
        • Tim:
          • There are versions of tales whether we record them or not
          • Checkpoint/tag: could be same as RR or not
          • RR: May mean that you have to create a checkpoint
          • Some RR may get published either through WT or to external repo
        • Kacper:
          • Create tale > tag > RR > publish
        • Kacper: Git/Github Analogy
          • Do two tags on local clone, push one to Zenodo (remote)
          • If you clone from Zenodo, you get one
          • But if you clone from WT source, you could get both of those tags
        • Mike H: Constrain
          • Before I publish, I do all sort of things that I don’t want to be published. I only want to publish the version of the code.
        • Tim: What is the “tale”?
          • Need to clarify
          • Tale-in-WT
            • Can have many tags/RRs
          • Copy-of-Tale-in-WT
          • Exported-Tale
          • Published-Tale-to-Zenodo
            • Zenodo has only the selected tag/RR
          • Tale-in-WT-imported-from-Zenodo

Updates

  • Kacper
  • Tommy
    • Currently testing RC3 (with firefox)
      • Ran into a few small non-blockers (need to see if they’re issues on chrome), issues yet to come
    • Whole Tale demo with an arctic data team, in general there was excitement; answered questions and solicited for Tales
    • Fixing bugs in dataone export
    • Planning to look at c2meta mid-week
  • Elias
    • Finished first implementation of Agave bdbag integration
      • Including refreshing token and updating Agave config and current files
    • Currently working on unit testing
      • Working on bag to test with
  • Craig
    • v0.8 PR review
    • UI/UX meeting
      • Github integration (needs clarification)
      • Analyze in WT
      • Reproducible Run is next priority
    • Reproducible Run write-up
      • Need to get info to Kristina, maybe?
    • Jetstream XRAS application
    • PresQT
      • Bags as primary interchange format – not bdbags or bagged ROs
      • API docs
        • Very early
    • Need to follow-up:
      • Website re-design discussion
      • RO-Crate test
      • Image preservation
      • C2Metadata meeting
  • Mike L
    • Further v0.8 testing
      • Unclear if there are further blockers for release
    • Angular rewrite
      • Still need to adjust Tale Workspaces modal to allow user to navigate down into Tale’s folders/files
      • Also trying to redesign/reimagine Tale Workspaces modal - current design could use some improvement
    • PR for “Currently running Tales” section
    • UI/UX discussions and clarifications (GitHub integration)
  • Mike H
  • Tim
    • Wondering what others have found out about C2Metadata.
    • Much of discussion in Friday provenance call was attempt to clarify for Bertram what the model of a Tale is in Whole Tale. One ‘reproducible run’ per Tale at a particular point in time (or tag), or more than one? Argued that a Tale represents (at a coherent checkpoint, not while things are in flux) the state of the Tale just following the ‘reproducible run’ that was just recorded.
    • Intrigued by idea (Mike H) that there might be a practical way to include in a Tale information that can be used to compare against Reproducible Runs at different tags.
    • Has anyone looked into Argument Graphs, Micropublication, and Nanopublications with respect to connecting artifacts with published papers?