What do we hope to acheive?

The first discussion was focussed on outlined the goals of ORSO.

Points for an ideal data format:

  • Priority listing: some information is more important to capture than other
  • Metadata
  • Cover both X-rays and neutrons, and facility and lab sources
  • Uncertainties and errors must be captured
  • The underlying format (xml/nexus) is not as important as the agreed data layout

Proliferation of software: currently many reflectometry analysis packages exist can we consolidate these somehow?

Education and communication is a vital aspect alongside the development of standards.

Standardisation: we aim to achieve some level of standardisation in the following areas.

  • Resolution and downstream handling
  • Units
  • What does data represent?
  • Where does generalisation end and specialisation begin?
  • Reproducibility
  • Guidelines: push for consistency with community and journals
  • Making sense of automated (in-line) analysis

How far should we go?

The second point was to discuss how many different things ORSO should comment/advise on.

  • It was generally agreed that specular reflectometry was within the remit, with possible investigation into off-specular in future
  • GiSAS could be a low barrier given the current work by the CanSAS community, however we do not want to duplicate effort
  • The NXcanSAS format would be useful starting point for a file format
  • Spin echo reflectometry is probabilty too much for now

Reduction – users should leave with analysable data (restrictions apply)

  • Instruments vary
  • Common ground but lots of variance
  • Raw and reduced data should be provided along with the reduction methodology

Industry involvement

  • The knowledge within X-ray reflectometry instrumentation industry would be valuable
  • Hopefully they will be able to contribute as well as take

Analysis

  • At least reduced data should be delivered
  • Analysis is a slope/incline
  • Software is often specialised for a reason, but there are common approaches
  • A shared library of these common approaches on some web resource would be useful
  • Benchmarking between the different softwares
  • Guidelines on unit tests
  • An analysis round robin
  • Analytical method comparison
  • Metadata regarding analysis
  • AI/ML in the future