Minutes Pencil Code Steering Committee: 23 April 2021, 13:00 - 15:30 CEST Present: Nils Haugen (chair), Piyali Chatterjee, Philippe Bourdin, Matthias Rheinhardt, Axel Brandenburg Agenda: - Approval of agenda - Election of secretary for the meeting - The upcoming PCUM - conference platform (break-out rooms) - anything else? - Should we use more special modules to slow down entropy increase? - is it needed? - which existing modules can be made "special"? - how to regulate new modules? - should we present some ideas to the PCUM? - Evaluation of the Pencil Code office hours - Travis - Code of Conduct (CoC) (version 1.4 vs. 2.0, not easy to find) - AOB ======================================================================= - Approval of agenda Under AOB, it was suggested to add: . feedback on the newsletter (Matthias) . the pc_varcontent.pro problem (Axel) . setting up the the key file for github . compile problem with io_dist.f90 Agenda was approved ----------------------------------------------------------------------- - Election of secretary for the meeting Secretary of the meeting: Axel volunteered and was approved ----------------------------------------------------------------------- - The upcoming PCUM - conference platform (break-out rooms) . Philippe reported on zoom tests done with Jennifer . We discussed usefulness of using Slack, is already in place . Discussed slowness of screensharing with zoom (Philippe mentioned dependence on window size) - anything else? . interest in a public participants list was expressed . several missing committers were identified Axel and Nils will contact several people ----------------------------------------------------------------------- - Should we use more special modules to slow down entropy increase? - is it needed? . increasing complexity of the code was emphasized . moving src/modules to special/ would not help . extra hooks required really complex modules (e.g., overset grids) . other issues: postprocessing ("cloning" of run.f90) . hidden operations (generation of io_dist.f90) can lead to problem (although they should never happen; item moved to AOB) . Philippe proposed a standardized interface (API) . Philippe will present this at the PCUM 2021 . limited availability only in idl (e.g. HDF5 conversion) is a problem - which existing modules can be made "special"? . only a few examples where mentioned: cosmicray*.f90, src/chiral.f90 . some modules could be moved to folders, e.g. magnetic/bfield.f90 . special.f90 has functionality to absorb dummy calls . Philippe demonstrated how Makefile.src needs to be adapted to do this also for energy_common using mkdummyinc. Nils to continue developing this further. . particle routines take up over 25% of all modules (could be moved to a particle/ folder) - how to regulate new modules? . was not discussed - should we present some ideas to the PCUM? . Philippe to present standardized interface (API) postprocessing idea . Matthias and Philippe to facilitate discussion on code complexity . Matthias and Philippe to discuss status on varcontent.pro problem ----------------------------------------------------------------------- - Evaluation of the Pencil Code office hours . We had 4 office hours so far (Jan, Feb, Mar, Apr) . Maybe not many people are aware of it and know what it is . examples: background on mass diffusion (not discussed in manual) . some information on code implementation to add to doc/implementation . to get supervisors to advertise to their students ----------------------------------------------------------------------- - Travis . Philippe to investigate and set up on travis.com . to check whether we have an account (travis.yml) ----------------------------------------------------------------------- - Code of Conduct (CoC) (version 1.4 vs. 2.0, not easy to find) . The version is 2.0, and is linked from README.md and from the Nordita homepage (http://pencil-code.nordita.org/contact.php) via https://www.nordita.org/~brandenb/pencil-code/PCSC/ ----------------------------------------------------------------------- - AOB - Newsletter as public outreach . good for archiving major steps . to collect and publicize feedback from students - github: to have access in future, we must in future generate a "Personal access tokens" (https://github.com/settings/tokens) For this to work, one must "Select scopes" for example "repo" and "Full control of private repositories" One finds this under "https://github.com/settings/profile" - beskow problem . the compile problem with io_dist.f90 could not reproduced. Nils had to leave at 15:15, after having discussed the Newsletter. Philippe had to leave at ~15:30 after having started access tokens.