2020-11-23: Development Meeting

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

Agenda

  • Updates
  • Big SDTL-ProvONE model change proposal – [name=Tommy]
    • Is the project about a ProvONE representation of SDTL? Or a hybrid model of both? (Mostly rhetorical. I’ll ask Matt J)
      • One of those approaches involves writing software that converts SDTL to RDF in accordance to the 4,000 line owl file.
    • Proposing that the ProvONE model has minimal sdtl
      • No extra objects like this or this or this
      • Stop trying to turn this into RDF
      • Embed data frame information in ports (original plan) which is the only SDTL included
  • Sharing - what to expect when two users edit the same tale – [name=Craig]
    • Tale
      • Editing keywords/description, etc
        • Won’t refresh in UI at all
      • We don’t expect two users to be editing the WT UI elements at the same time.
    • My experience w/ workspace:
      • long lag between edits: 10 second webdav
      • File created by user2 never appears for user1
      • User2 cannot edit or remove file created by user1
    • Should we lock it?
      • Tim: Lock/Unlock request?
      • KK: Analogy of project space in HPC cluster
      • CW: What are we saying: We allow simultaneous editing (ala HPC/shared system)?
      • KK: This is an edge case.
        • The more interesting is read-only with private work. Running classes.
        • /etc/davfs2/davfs2.conf

Updates

  • Tommy
    • Uploading the results of BD+C2Metadata to DataONE
      • Next Steps
        • Bring in ProvONE/SDTL Converter
        • Insert RDF into Resource Map
    • SDTL model meeting in probably 2 weeks (take 20)
      • I’m happy with the way I’m modeling ProvONE stuff, really have no idea about SDTL
      • Will also be updating the provone-sdtl documentation so that it reflects the current model
      • Also going to propose a big change
  • Craig
    • MATLAB demo video for Mathworks folks
    • Requested annual licenses from Stata
    • PR review and testing
  • Kacper
    • PR review and testing
  • Mike H
    • PRs updating versioning addressing reported issues
  • Tim
    • Outline for talk on ProvONE at workshop next week:
      • Title: ProvONE – A Conceptual Model and Vocabulary for Prospective Provenance
        • Overview – Provide a context for understanding ProvONE
        • Role of provenance in scientific research
        • History of provenance models and standards
        • Origin of ProvONE
        • Understanding its capabilities and limitations
        • Where we go from here
      • How scientists think about provenance: it’s not meta
        • ChIP-chip workflow diagram – made by genomics researcher
        • Note two kinds of boxes—programs and data
        • Data flow is the natural model of provenance for these researchers
        • The answer to the question “how was this result obtained” is answered by a dataflow graph culminating in the result
        • More specific questions about the provenance of results are queries on this dataflow graph.
      • Scientific workflow system – automating what researchers already do
      • Scientific workflow systems and models of computation.
      • The process network model of computation (Kepler/Ptolemy-II terminology)
        • Workflows, Actors, Ports, Channels, Sub-workflows
      • Early models of scientific provenance – RWS and COMAD
      • The PROV model – retrospective provenance
      • The ProvONE model – prospective provenance
      • Dude, where’s my data?
      • YesWorkflow – Prospective provenance for scripts
      • YesWorkflow – Reconstructing retrospective provenance
      • Queries – automating what researchers already do
      • Building on PROV and ProvONE
      • Towards comprehensible provenance
  • Mike L
    • Added a “Shared with Me” tab to tale-sharing PR
    • Address PR review comments and review open PRs