Try   HackMD

NICEST2 final day-3 morning

Link to the shared document https://hackmd.io/@nicest2/SyZlj6HI2

  • What we learned about FAIR (Link to presentation: https://docs.google.com/presentation/d/1gjobLBZxN1fdOh6F8lXn6sROq-6DDgo4VRsK-EXz7RU/edit?usp=sharing)

    • Hackathons & workshops (M4M with GoFAIR)

    • EDAM onthology (with Matus Kalas)

    • The word "data" is used a lot in the DoW

      • "Research Object" is a wider concept
      • Lack of governance/authoritative body
        • Who "decided" what metada is needed, which software is to be used, etc.?
        • Until that is sorted out there are only local "solutions" implemented
          • Interoperability?
          • Roles and responsibilities?
          • Incentivization
          • Efforts vs. payoff?
          • Hamish believes this has to be a top-down approach
          • Oskar realized that there are benefits in the long run to make things reproducible for one´s future self
      • Need to engage with the community
      • Reset the goals (beyond data)
    • These are many points that we overlooked at the beginning of NICEST2

      • Really overwhelming at the begining for reseachers
      • Allocate funds for data stewarts or the like to help <- this will be easier if the incentive comes from above
    • For the afternoon HORIZON-INFRA-2024-EOSC-01-01 discussion

      • SND (https://snd.gu.se/en/contact/contact-management) feels like they would be a good lead partner from Sweden - if they are interested.
      • NAISS (the new SNIC) may be able to act as a back end resource provider
      • Researcher/User engagement - should be an important WP. Workshops/meetings/etc. and promoting early adoptor examples & use cases.
  • ESGF in the Nordics

    • Lots of activities carried out under NICEST2 (CSC-ESFG node, NorESG/NIRD, assisted DMI, indexed data, publishing tool with custom scripts folowing strict rules & procedures -> automated workflows to minimize human errors)
      • Precheck module (supports SLURM) uses containerized deployments of nctime (to ensure that data published has the correct "time bounds" and "time axis aquareness", cmorfixer, cdf2cim, etc.
        • Not allowed to delete data (keep metadata)
        • Creates a Makefile that can be executed on the ESGF datanode
        • Thredds dataserver catalog files
        • Spotted issues with duplicate files
    • Demo