Tuesday, 8 December, 14:00 UTC
Present: WG, SC, RO, BW, MK, FA, RB, PJ, PC, HG, SB, TM, MO
The resolutions from NIAC2020 were all passed
BW will check that tickets related to these decisions are closed. Any issues that are remaining, were not discussed, or inconclusive, will be handled at a future NIAC
RB had previously mentioned a program he has developed for generating example NeXus files from NXDL. He described how this had been recently extended to also generate example h5py files, which all agreed was a useful addition.
The recent code camp was discussed and it was agreed that the event had been good at focussing effort and making progress and so another virtual code camp would be arranged for 2021. BW will create a poll to choose a preferred date, sometime in february/march/april. We will discuss code camp agenda items at next TC.
PJ mentioned the recent GitHub change of default master -> main branch. He will investigate the impact of making this change over the next month or two.
SC mentioned the recent policy change at Travis CI and how this was causing problems, PJ suggested we move to GitHiub actions. He shared a couple of useful links on migrating travis to githuib actions and pythin testing WG also shared an example of how this had been done in another project
SC has been working on including nexus in Linux package repositories and it is now available in Fedora linux
PJ is continuing to look at the generation of html anchors, which are related to persistent identifiers (PIDs) proposal presented by HG at NIAC2020
The sphinx version of the user manual seemed to have stopped building on the STFC server. Action FA to check (turns out to be missing python package after O/S upgrade; now fixed)
The feedback from the questionaire circulated to the mailing list was dicussed. Most facilities are either no longer using NAPI, or are building their own copy and able to maintain that themselves. A few people are still depending on a maintained NAPI, namely the Mantid and McStas software and a planned upgrade at the LANL facility. The bindings being used here were HDF4 and HDF5, nobody reported use of NAPI XML bindings.
It was decided to keep NAPI in maintenance mode, but to make a few changes to ease workload:
RO suggested that NAPI could be moved into a separate repository and community managed separate to nexusformat. As we are keeping minimal maintenance, it was decided to keep it in its current location for now, but the situation would be reviewed. Separating it from the main nexusformat organisation would make it easier for other interested parties to maintain e.g. otherwise dropped formats/utilities/bindings as needed, they could currently fork the existing code and do this, but that may lead to confusion between “official” and “unofficial” bindings.
WG mentioned that ORNL were looking at ways to improve the interaction of programs with HDF5 in a multi threaded environment.
Proposed for mid - end January 2021 (poll)