VLA Configurations. LorantSjouwerman did not get a chance to think about a better method but will do so in the next week.
VLA Widar. We discussed the proposal from LorantSjouwerman on how to configure Widar. The main difference compared to how the OPT currently works is to create the setup in the time independent Barrycentric frame first and then apply the Doppler corrections at validation for a given time. Here are the steps:
Lines: specify rest frequencies
Line Placement: create Barycentric line subbands. Software will determine the observing band, suggest 3-bit or 8-bit, and place the sub-bands. The user can override these values.
Sub-bands: modify sub-bands and add continuum.
Doppler: specify Doppler position.
Doppler: Doppler switch.
Basics: choose integration time (guess from configuration/band).
VLBA. Could the same approach used for the VLA Widar correlator be used for the VLBA correlator? MarkClaussen is skeptical that this would be easy to do. AmyMioduszewski thinks this is doable but the constraints would have to be worked out. For example, there are constraints with the LO as well as other components. This is currently not done in the PST but Sched does check this and appears to be reliable. Currently, VLBA users typically read the documentation to figure out if what they are proposing is plaussible and then find out when using Sched.
GBT. Could the same approach used for the VLA Widar correlator be used for the VLBA VEGAS correlator? RyanLynch thinks this could be done. We might also consider how mapping would be specified. The user input necessary to specify a map is not too different than for a VLA mosaic. A field position and a field of view. There was some concern about allowing the software to suggest how best to produce the map (e.g., pointed versus OTF); a better approach might be to give the user options to select. ToneyMinter noted that there are often private backends (e.g, FLAG). But the GBT is not in scope for SRDPs, though it should not be ruled out, and therefore we can provide a textbox if necessary for the user to specify any details (as we do now).
Overview. We agreed that we really do not need both a novice and an expert mode. That is, the proposal above should work for both. At each step the user is allowed to override the decisions. We also agreed that the best approach would be to have only high level validation at each step (e.g., this field requires a decimal), and that once the steps were completed that a low level validation would be performed to make sure the configuration was possible. We agreed that the detailed requirements that provides the algorithm to configure the correlator would be in another document.
LorantSjouwerman will think about more detailed requirements using the intent to specify the VLA configuration(s).
DanaBalser will update the document by Monday (3 December) so people can read this before the next meeting.