Minutes of SunPy Board Meeting 20180221

Date of Meeting

21 February 2018

Attendance

• Present: David, Stuart, Juan, Steve, Monica, Sabrina, Jack, Kevin, Russell
• Absent: Albert

Agenda

1. Releases update
2. GSOC update
3. SPD committee update
4. NumFOCUS survey about shared infrastructure
5. IRISPy - news regarding relations with IRIS team?
6. Definition and process for affiliated packages (including provisional status)

Meeting Notes

Topic 1:

• 2 new releases: 0.8.4 & 0.8.3

  • 0.8.4 required for AstroPy compatibility
  • 0.8.3 does not currently work with AstroPy due to an issue with a bug fix

• In progress: 0.9 & 1.0

  • 0.9 includes change from current library concerning VSO interaction & will include a revised JSOC client
  • 1.0 deletes Python 2 support and uses AstroPy time object instead of date time object
    • http://docs.astropy.org/en/stable/time/

New VSO/SunPy liaison

Topic 2:

• GSOC progressing well
  • Currently 5 SunPy projects; open to more
    • http://openastronomy.org/gsoc/gsoc2018/#/projects
  • Looking for more mentors
    • https://google.github.io/gsocguides/mentor
    • https://github.com/OpenAstronomy/GSoC/blob/master/application_students.md
  • Tutorials:
    • http://openastronomy.org/gsoc/gsoc2018/#/projects?project=astropy_learn_website

Topic 3:

• SPD committee agenda item:

  • Jack will present request to SPD committee at upcoming TESS meeting

• TESS presentations:

  • SunPy poster submitted to TESS meeting via Andrew Inglis and Jack
  • Python NDCubed poster submitted by Danny Ryan

Topic 4:

• NumFOCUS survey
  • Questions about what support is needed from NumFOCUS (e.g., web hosting)
  • Discussion about possibly requesting support for FIDO upgrades
  • https://www.surveymonkey.com/r/shared_infrastructure

Topic 5:

• IRISPy talks ongoing
  • merging of parallel IRIS and SunPy efforts led to development of splinter NDCubed effort led by Danny Ryan and Stuart
  • GUI QuickLook tool guidance being managed by Stuart
  • NDCubed is a generic tool that will be merged into core package
  • IRISPy is not far enough along yet to be an affiliated package

Topic 6:

• Affiliated package (AP) management discussion
  • discussion over instrument specific packages
    • core packages built to work on already calibrated with minimal processing
    • affiliated instrument packages handle data nuances
    • https://github.com/sunpy/sunpy-SEP/blob/master/SEP-0004.md
  • qualifications to be listed as an AP need to be delineated
    • tests/documentation/use of available functions
    • set of enforceable standards needed
    • using AstroPy experience as a foundation
      • official registry
      • tiers: provisional, accepted[, sponsored/branded]
      • light code review process
      • http://www.astropy.org/affiliated/
      • https://github.com/ejeschke/ginga/issues/555
  • purpose of APs to not duplicate code, to get widespread use, and to work together for the positive evolution of the tools
    • build with core library
  • implementation more difficult than expected based on IRISPy experience (e.g. GitHub use, ownership issues)
  • voting process for accepting APs under review, possibly separate from the board

AOB