Telco 20200904

Date

Friday, 4 September, 15:00 CEST (+0200)

NIAC2020
Github questions
NXmx “Gold Standard” changes:
- definition changes
- cnxvalidate pull request
Mailing lists:
- note that signing up for the NIAC list, doesn’t automatically add you to the Nexus discussion list.
- Should we remove people from the NIAC list if their membership of the NIAC ends?

Present: BW, HB, PJ, RB, HG, VAS, PC, WF, AP, MK

Pull request is making sure the tags that the community needs are included.
HB asks that people take a look and he hopes that the PR can be accepted by September 9th.
HB says that Dectris would appreciate this being sorted out very soon so that they can finalise their software.
BW promises to look over it during the weekend.
Replacing _error with _errors generally needs its own github issue to make sure the community can see why the change happened (to fit previous NIAC decision).
We should break them up into separate PR to tackle it piece by piece.
Dectris is specifically asking for a NeXus release to quickly follow the NXmx pull request.
- We have a milestone for this aiming for the NIAC in October.
VAS offered to give a pull request to have CMake find the HDF5 library from an environment variable.

VAS says NXcxi_ptycho is restricted to to raster scanning.
BW recommends NXstxm as it is quite flexible.
The inconsistent naming issue should be brought to the NIAC.
- we should avoid ambiguity!
PC to look into who at Diamond might be using NXarpes.
Russ to make a PR to suggest appropriate changes.
Should make sure that the NXmx PR #793 contains a fix for issue #782

BW points out some confusion about the mailing lists and suggest that we add a note that subscribing to one list doesn’t automatically subscribe you to the other.
VAS suggests a purge is required to stop it from becoming a public list
HB and WF suggest not pushing people away - perhaps just ask if they want to be removed

PJ says polling on github might not work.
VAS suggests doing a simple thumbs up/down on the github issue to indicate their vote.
BW suggests that emojis might not be best because it is easy to mess up - but VAS corrects that a second click erases the mistake
BW and VAS propose emoji voting:
- Thumbs Up for Yes,
- Thumbs Down for No,
- any other emoji for Abstain.

RB asks about the code camp - BW points to the page
RB brings up syntax issues - BW suggests making single PR to address each type of issue for all appl defs at once.
RB says links are tricky to name automatically and there are lots of inconsistencies
- PJ says part of the problem is that some names need to be flexible and we decided to try and indicate flexible names as upper case, while required names are lower case.
- VAS suggests that cnxvalidate doesn’t check what kind of object a “link” is, just that the object is present.
- HB says the problem is going to need to be figured out for when virtual datasets become common.
- MK says the target_path of a link is somewhat historical and maybe we should review that usage.
- VAS says ESRF is using these links
- BW suggests we make this an issue to discuss in the code camp.
- HB points out that HDF5 has changed a lot in the past few years.
Next telco will aim for the last week of Sept or first week of October. Poll