Date: Mon, 10 Apr 2006 13:30:25 -0600 (MDT) From: Walter Brisken Subject: Comments on sections 1-3 in UIUG report Hi Steve, I'm looking through your draft report. It looks pretty good. Thanks for taking the initiative to rank the options. Here are some comments on the first 3 sections. More later... -W UI-1-R10 -- I think the CLPI already will be an editor capable of editing the SAVEd parameters. If you intend to allow text editing of these files then that should be made explicit. UI-1-R20 -- This deserved far higher priority. regarding UI-1-R21 -- Should we have a Timescale '@' (ie the ASCII character preceding 'A' for a CLPI mockup or a final CLPI with partial functionality for the users to work with before making a final decision here? I think under UI-1 you need to add some form of minimal-match / tab-completion / (non)case-sensitivity with priority P:2 T:A regarding UI-1-R26 -- should we add another P:3 T:E requirement that the GPI can mirror an active CLI with GPI parameters updated as entered at CLI? UI-2-R8.9 --I think given UI-2-R8.4 this implies that a DBCON-like functionality also is part of this. For example, combining A, B array data and selecting only the (calibrated) target source and perhaps some spectral averaging could all be done in this operation. Given the complexities implied by UI-2-R9, such tasks may all benefit from a "STATE" command (to be optionally executed before "GO") which would, in English, describe the operation to be done, such as: STATE About to image mesasurement set = AB911.ms Source = 'J1234+56' Field ID(s) = 1,2,3 Algorithm = 'Clark Clean' Using channels 3-11 from ms (9 total) spanning 1451.1 to 1452.1 MHz Total visibilities to be gridded = 123523 Producing image cube with 12.2" x 12.2" resolution image size = 1642 x 1921 pixels (5.6 x 6.5 degrees) 3 output channels spanning 1451.1 to 1452.1 MHz 2 polarizations = RR, LL Image center = 12h34m56s +56d45'12" (J2000) Cleaning will stop at residual flux density of 0.1 mJy This could issue various warnings that the user can choose to ignore: WARNING: Time-average smearing will be significant beyond 1.4 degrees radius. reduce time averaging to or reduce image size to by to prevent this WARNING: Wide-field imaging options should be used WARNING: Inusfficient resolution to avoid cleaning artifacts. reduce cellsize to (, ) or uvmax to to prevent this This sort of functionality would require access to the ms. UI-2-R10 -- also for combining datasets, especially from different arrays/configurations. Sec 3.3 -- migration to common/consistent parameter names will help quite a bit here. Should there be a requirement for "HELP " or "EXPLAIN " as in AIPS? This should be context sensitive (ie general outside of a CLPI and task/function specific within). =============================================================================== Date: Tue, 11 Apr 2006 09:38:42 -0600 (MDT) From: Steven T. Myers Subject: Re: Comments on sections 1-3 in UIUG report Thanks for the comments, I'll incorporate these into a new draft. ------------------------------------------------------------------------- On Mon, 10 Apr 2006, Walter Brisken wrote: > UI-1-R10 -- I think the CLPI already will be an editor capable of editing > the SAVEd parameters. If you intend to allow text editing of these files > then that should be made explicit. Yes, thats whats intended. I will make this clear. > UI-1-R20 -- This deserved far higher priority. Its not clear how essential this is, hence low priority (its potentially useful, but not critical, while there are many things that are critical). Note many tasks will be in C++, and thus this would not be applicable. For Python tasks, this would require some sort of interpreter to unravel logic (unless you just want a dump of the script). > regarding UI-1-R21 -- Should we have a Timescale '@' (ie the ASCII > character preceding 'A' for a CLPI mockup or a final CLPI with partial > functionality for the users to work with before making a final decision > here? I thought about that, maybe "A+" for stuff we want to see mockups of asap? > I think under UI-1 you need to add some form of minimal-match / > tab-completion / (non)case-sensitivity with priority P:2 T:A Good - I forgot that one! > regarding UI-1-R26 -- should we add another P:3 T:E requirement that the > GPI can mirror an active CLI with GPI parameters updated as entered at > CLI? I would say this is too complicated with limited usefulness, but if others concur. > UI-2-R8.9 --I think given UI-2-R8.4 this implies that a DBCON-like > functionality also is part of this. For example, combining A, B array > data and selecting only the (calibrated) target source and perhaps some > spectral averaging could all be done in this operation. Concatenation is (already) an option covered by SSR and thus needs no recommendations here. > Given the complexities implied by UI-2-R9, such tasks may all benefit from > a "STATE" command (to be optionally executed before "GO") which would, in > English, describe the operation to be done, such as: > > STATE > > About to image mesasurement set = AB911.ms > > Source = 'J1234+56' > Field ID(s) = 1,2,3 > Algorithm = 'Clark Clean' > > Using channels 3-11 from ms (9 total) > spanning 1451.1 to 1452.1 MHz > > Total visibilities to be gridded = 123523 > > Producing image cube with 12.2" x 12.2" resolution > image size = 1642 x 1921 pixels (5.6 x 6.5 degrees) > 3 output channels spanning 1451.1 to 1452.1 MHz > 2 polarizations = RR, LL > Image center = 12h34m56s +56d45'12" (J2000) > > Cleaning will stop at residual flux density of 0.1 mJy > > > This could issue various warnings that the user can choose to ignore: > > WARNING: > Time-average smearing will be significant beyond 1.4 degrees radius. > reduce time averaging to or reduce image size to by > to prevent this > WARNING: > Wide-field imaging options should be used > WARNING: > Inusfficient resolution to avoid cleaning artifacts. reduce cellsize to > (, ) or uvmax to to prevent this > > This sort of functionality would require access to the ms. This is a good idea in general - I'll add getting understandable info like this as a separate rec. > UI-2-R10 -- also for combining datasets, especially from different > arrays/configurations. I'm not sure what you mean here. R10 is for calibration of weights (DOCALIB=2 in AIPS) so that they are in the correct units (1/sig^2 normally) etc. Setting weights between datasets is handled under some sort of imaging weight control (in the appropriate task). Do we want a rec on that here? > Sec 3.3 -- migration to common/consistent parameter names will help quite > a bit here. Yes. > Should there be a requirement for "HELP " or "EXPLAIN > " as in AIPS? This should be context sensitive (ie > general outside of a CLPI and task/function specific within). Yes - I'll add that! ===============================================================================