Telco 20160531¶
Date¶
Tuesday, 31st May, 16:30 Central European Summer Time (Copenhagen)
Hangout Link: https://plus.google.com/hangouts/_/j72qwlvegiojjpt3a36pfhow5ua
Agenda¶
Welcome and agree Agenda
registering for NIAC and potential hotel shortage
github issue of the day - suggestions welcome (there is a github label now)
–definitions–
#387- proposition for version number of NeXus definitions
–code–
329 - can we remove the issue branch from the repository?
419 - still waiting on a response from Stuart where to install the header files
425 - done
426 - Maybe Tobias can have a look at it
420/421 - currently working on it (this work should be continued by the original author - the code is a total mess)
release NAPI/tools with a fix for https://github.com/nexusformat/code/issues/414 - what else needs to go in?
review and discuss accepting NXcontainer pull request: https://github.com/nexusformat/definitions/pull/472
Intent to present NXcanSAS for ratification as application definition by end of June
items for NIAC
CXI nexus integration question about virutal dataset
Is a new graph of the base class hierarchy useful? Used to be in the manual.
Next meeting
AOB
Minutes¶
Present: HJB, MB, PJ, MK, EW, TSR, AB, ORNL (Mathieu and Steven)
Welcome¶
Especially the ORNL!
Github Issues¶
definitions #387 - consensus is that this may solve one problem well, but certainly does not address all aspects of versioning. There was debate whether versioning the definitions as a whole was actually required. We encourage people to look into this to have an informed discussion at the NIAC and code camp.
code #329 - branch can go
code #419 header files will be added
code #426 TSR’s problem, but adding an API function to expose information about the backend may be required
code #421 assigned to different developer
NAPI release update¶
Could be done from master this week with the limited amount of changes that went in so far. The release with support for HDF 1.10 may be NAPI 4.5 anyway, due to the high number of changes.
NXcontainer¶
To be merged. There is a low bar for accepting contributed definitions.
MB mentions that the use of NXshape in NXontainer may require revision. That is to be evaluated and presented at NIAC.
Items for NIAC¶
There is a repo for NIAC items. Please add thing that will need in depth discussions early so we can be prepared. We may assign items to people for preparation in later telcos.
CXI Compatibility¶
CXI uses a n by 3 array where NeXus is more flexible and allows multiple 1D arrays of length n. To be compatible with both standards for a smooth transition having the data both ways may a way. Doing that using HDF virtual dataset would save storing the data twice. As far as NXDL is concerned that data just needs to be provided twice, virtual dataset is transparent to our abstraction.
Class diagram¶
Very informative, might be too scary for newcomers. But we will try a nicer layout. It is a useful tool for us in any case. Could be in an advance section of the manual.