Thursday, 1st October, 8am PST, (5pm CEST, 4pm BST, 11am NY, 10am IL+TN)
Present: HJB, EW, MK, MB, TSR, PJ, AB
ACTION: Everyone register, EW to send reminder to mailing list
EW had a look and found no problems (so far)
MB has no problems with the current proposal, but would like to add the use case of recording the scan intend. One way of doing that would be to use the old NeXus “target” linking scheme to point to an NXpositioner or so. He’ll generate a proposal.
HJB leaves the meeting due to conflicting commitments.
Pete clarified that the attempted old solution didn’t lead anywhere promising. So having Freddie looking at this with fresh eyes is a good idea.
Mantid does require not only positions of elements on the beamline, but also their shapes. For visualisation but also modelling and simulations. In the discussion it became clear that knowledge about the adopted CIF geometry is not as widespread as we would have liked. (See ticket 397.) Tobias will add some information there.
The existing NXgeometry or NXshape were not suggested as good starting points. Either working from what mantid currently uses, or starting from some general purpose geometry package would be reasonable. Avoiding the re-inventions of wheels for object placement.
Will put of a reduced feature set release soon and ask for volunteer effort to help with things further down our priority list.
Mark B mumbled something, could make out some words no coherent meaning. ACTION: MB get better mike.
MB mentioned very careful background subtract schemes with half sample containers in the beam (upstream and downstream) and the requirement to provide more information on the holder material. He will propose something and potentially provide a drawing if helpful.