Abstract

Quick & dirty document to kick off design of user interface for hardware configuration for EVLA. Beware: this is just a brain dump with little organization or coherence.

Introduction

It's time to create visual specifications for how we're going to present hardware configuration to observers. The layouts presented herein are not proposals for completed designs, but are meant to discover what features the observatory staff things we should and should not have in the user interface. Use them as seeds for futher ideas. The first place to look is at the WIDAR document: Memo #012. For now we're interested mainly in the spectral configuration portions of that document.

While i had originally wanted to take the approach that our UI would completely mask the specifics of the EVLA, i no longer think this is wise. To begin with, all attendees of the weekly SSS mtgs feel it is a good idea to have the users choose the receiver band(s) of interest, so that's our first exposure of the instrument. I think it is also wise to do something that exposes, though perhaps in disguised form, the fact that we have LO tuning restrictions and restrictions related to the correlator (esp. the 128MHz boundaries).

The next couple sections contain roughed-out UIs for discussion. I'm not fond of drawing, so i left a lot of things out of the dgms that need explanations. Reduced-sized versions of those layouts are given below. Clicking on the picture will bring up a full-sized image in a new window (or tab). It may be useful to have the full sized drawing up for the discussion.

Coarse & Fine Tuning Diagram

This section pertains to the first dgm (in new window). The goal of this dgm was to allow the software to think in terms of the LO/IF vs correlator split, without specifically telling the users about that split.

Coarse Tuning

The first thing the user does is select the receiver band. A default selection will appear initially. Unlike what is shown, all buttons would have the freq range. The selected band is highlighted and causes the section headers ("Coarse Tuning..." and "Fine Tuning...") to be updated.

The next section deals w/ the LO/IF in disguised form. This section is split into two subsections, one on top of the other, separated by a frequency scale. In each section the user may make a choice about the number of bits to sample. This choice impacts what is shown in its subsection. In the dgm, the top subsection uses 3-bit sampling, leading to two pairs of basebands, while the bottom subsection uses 8-bit, leading to only one pair of BBs. From the user's standpoint, the A1, C1, etc., labeling is not meaningful and serves only to tie the coarse and fine tuning displays. For NRAO staff, however, these monikers carry physical meaning. The RCP/LCP, on the other hand, should carry meaning to the user. Each pair of BBs can be moved by the user along its horizontal line. How far left or right it may be moved will be dictated by the underlying LO/IF code.

The arrow on the extreme right of the coarse tuning section might open up an expert panel. This panel could hold user-settable frequencies for all the LOs in the system. Changing these values would move the graphical BBs.

Though not shown, spectral lines and RFI could be displayed in the Coarse Tuning section.

The BB pairs would probably be clickable and allow direct entry of the starting frequency for that pair. Direct entry of the frequency will often be more useful than aligning on a particular pixel.

Fine Tuning

For each BB pair in Coarse Tuning, we have a display in Fine Tuning that is color matched. It is in this section that users place their subbands. Look at the red A1/C1 portion of the dgm. The vertical gray lines represent the 128MHz boundaries that a single SB may not straddle. We could show spectral lines and RFI here. Just below C1 would be a frequency or wavelength scale (user convertible, i s'pose). Users would be allowed to specify subband-pairs and single-polarization subbands, as shown.

By moving the BBs in the coarse tuning section, ALL the SBs for those BBs would move in sync. Visually you wouldn't see the SBs move at all. Instead it would look like the frequency scale is moving. A user would be able to select one or more subbands and move them relative to each other. We should discuss, though, whether or not they could move them outside the bounds of the baseband. If we were to allow it, we would need to have such a movement retune the BB. We'd also have to worry about the subband on the other end of the BB -- it might now be outside the BB. Restricting the movement of SBs to the current BB settings would make those issues moot. We could allow users to pan this display beyond the BB so that they could see if they are just missing any lines of interest.

Not shown, but needed, are icons for adding, deleting, selecting, etc. the individual subbands. We could repeat these icons for each BB, or we could make one set and have it operate on the "active" subsection of the Fine Tuning section.

Missing Items

Also not shown is the correlator resource diagram. We could put this at the bottom and have it visible at all times, or we could hide it somewhere and make it visible upon request. LO settings could be treated likewise.

Four Spectral Regions Diagram

This section pertains to the second dgm (in new window). As before, the goal of this dgm was to allow the software to think in terms of the LO/IF vs correlator split, without specifically telling the users about that split.

Missing Items

I forgot to show receiver selection. I think this would be a bar similar to what is in the 1st diagram. Users would make a receiver selection, or accept the default, before doing anything else.

Bottom of Diagram

I think we could turn this to an "experts'" section that is initially hidden, but made visibile upon request. It would contain the correlator resource diagram, LO tunings (that can be modified), and the 3-bit / 8-bit choice.

Spectral Displays

There is one rectangular box for each baseband. We could either: a) display only the active basebands (there would be 2, 3, or 4) or b) always display 4 but disable input on the inactive bands. The colors used as borders for the spectral displays don't tie to anything and could all be made the same.

The horizontal line running through the middle of the frequency window is used to distinguish subbands that are RCP only from those that are LCP only. Subband pairs straddle both sides of this line. The gray vertical lines are the 128 MHz boundaries. The green vertical lines are absorption / emission lines. The gray dashed line and colored area is RFI. The black vertical lines and gray areas to the outside of those two lines represent the edges of the basebands. No subbbands may enter those areas. Users could turn RFI display on or off. Same for spectral lines.

Icons are needed for adding, deleting, etc., subbands and lines. We could repeat such icons in each of the BBs, or we could merge the rightmost section of the 4 BBs into one and have the icons work only on the "active" BB.

Users would be able to pan and zoom the frequency windows.

Users could set the frequency of the leftmost SB (that is, set the starting freq of the baseband) by direct entry, just to the left of the frequency window. This value, though, would be restricted by the range of the selected receiver band. Changing this frequency would change the start of the BB, but would keep all SBs right where they are in relation to the start of the BB. Visually, this would look like the frequency scale, spectral lines, and RFI graphs all moved while the SBs stayed in place.

I don't remember what "other useful things" are.