Minutes from SSS Mtg 2008-Jul-22

Attendees: B.Butler, D.Harland,  B.Hesman, G.v.Moorsel,
           F.Owen,   J.Rochford, B.Truitt, S.Witz

In this meeting we gave summaries of our status on the items outlined in the report written by Lorant, Gustaaf, and Mark regarding OPT readiness for Ka, and on the goals for the quarter. At the end of the meeting we discussed the advanced search for calibrators.

* * *

Items from the OPT Readiness Report

A. User Help & Documentation
JIRA: EVL-228, EVL-420, and some others.

[BT] Lorant plans to write whole document at once and then go back and add tags in the document for flyover help, etc. We will invent / change keys as directed by Lorant.

[DH] We are not working on this topic; the ball is in Lorant's court.

B. Stable & Robust Web Interface...
We think we understand what is meant here -- the application shouldn't crash, but if it does, we should be able to restore the users' data to where it was prior to the crash.

We have noticed that the OPT performed better on webtest after we shut down non-SSS applications. We are not, though, doing any work right now with respect to web/application server tuning or troubleshooting.

James completed EVL-583, so now the RCT will copy changed catalogs on "undisciplined exits".

[JR] If the users' browser crashes, but the user logs back in before the time-out period (4 hrs) expires, the state of the tool should be as it was just prior to the crash. If the user logs back in after the time-out expired, the system will make copies of all changed catalogs and save the unsaved changes in those copies, but leave the originals as they were. This has been done for the RCT catalogs, the SCT catalogs, and the projects of the OPT.

C. Single Sign-On
Stephan tried some open source software (CAS) that addresses this issue and was able to get it to work. At this point, this was only a test. The hope is still to have E2E take care of this NRAO-wide issue, but nothing will happen there until the new fiscal year, at the earliest. Plan B is to take Stephan's work and put it in place for the OPT-related applications.

D. Generation of Readable Summary Listings
JIRA: EVL-43.

Dave took an excerpt from Lorant's summary of last week's meeting and attached it to the issue. If necessary, Brian will ask questions about the req'ts via JIRA.

E. Improved Calibrator Searching
JIRA: EVL-300, EVL-301.

Dave took an excerpt from Lorant's summary of last week's meeting and attached it to EVL-301.

First we had a small discussion on the desire to use VLA-catalog-specific search parameters (calibrator codes for a given band and array configuration) in our more generalized tool. We acknowledged the impurity of using these items, but also acknowledged the practicality of using them. A longer discussion on the advanced search form took place at the end of the meeting. See the end of this note for a recap.

F. Writing Correct Observing Scripts...
JIRA: EVL-445.

This is about 95% done. Some to-do items:

  1. The scheduling block needs a start-time property.
  2. Error messages need some work.
  3. Biggest item is tipping scans, if we need to do this at all.
    BB: Send Claire an email, waiting for response. Assume that we need to do this, but do it in the simplest possible way. User may choose to do a tip, can specify azimuth, up/down vs down/up, but not the elevations. They will get a std 7-point tip. Do simplest possible thing in model2script.

LO/IF setup is done.
Secondary pointing is done.
Delay calibration is done.

G. Implementation of Doppler Tracking
JIRA: EVL-675.

Dave took the info from Lorant's summary of last weeks' meeting and created issues EVL-685 and 686.

Dave has outstanding question to Bryan re: use of SLALIB for calculating velocity of source relative to position on earth. See EVL-677.

Lorant's summary mentions a "FIN" card. In VLA documentation i saw something called an "FIS", or "FI", card for "fine tuning" (FI Card). Is this the same thing?

[BB] It's an FI card; the 3rd character is a modifier. N=No change, S=set.

Note: The new button to be added to the VLA cfg s/b doppler vs not doppler, not sky vs rest frequency.

Discussion of "FIN": Let there be two consecutive scans in a scheduling block. The first scan is for a target source and uses a hardware setup that says Doppler tracking should be performed. The second scan is for a calibrator. The FIN concept is not the same as simple using the same hardware configuration that was used in the prior scan. Instead it is a signal to not do any changing of the hardware at all -- specifically not retuning the LOs.

To support this we will need some new control on the scan page to let a user specify this behavior. In such an instance the user should probably be prevented from selecting a hardware configuration. This will have ramifications for the XML schema (resource needs to be optional) and for scan validation (resource to be null if and only if this new option is on).

H. Capability to Specify Dynamic Constraints
JIRA: EVL-682.

Dave took req'ts from Bryan's original xml schema and created EnvironmentalConstraints class. Brian exposed two of these (wind velocity & API) in OPT.

Need to know what to do w/ these re: automated scheduling.

[BT] Scheduling summary doesn't deal with solar separation yet.
[DH] I have an outstanding email to Bryan on this.

* * *

SSS 2008.Q3 Goals

1. Produce scripts from the OPT...
See F, above.

2. Support 10-stn prototype WIDAR
Dave did some work that involved incorporating VCI schema generated classes into the SSS body of work. Some rearrangement of the SSS code was necessary. During and after the VCI review last week, though, it became apparent that some of the VCI will be changing. Dave would like to see it finalized before doing too much more work against it.

3. Support Doppler tracking
See G, above.

4. Allow users to enter scheduling constraints in OPT
See H, above.

5. E2E user DB...
See C, above.

6. Advanced source searching
See E, above.

7. Improve process for submitting EVLA Observe files
[SW] We gathered list of what operators do. We don't yet understand the "why" of each of the things they do, but we are making some slow progress.

8. Procedures for software versioning & production software updates
[SW] I'm using parminator to do some experiments with tagging and branching in subversion. Kicking around ideas for using maven profiles for the data access layer. We might be able to incorporate these into a single DAO jar and trigger the profile at runtime, as opposed to triggering at compile time and making different DAO jars.

[BT] Subversion has a python script, svnmerge, on their website that helps with the bookkeeping associated with merging branches.

* * *

Discussion on Advanced Search for Calibrators

(Write-up provided by BT.)

A paper mock up of the new search interface was reviewed. In summary: "Quick Search" and "Sesame Search" boxes will be added to the top of the left pane (above the tree). The current search boxes will be removed. An advanced search node will be added as the top node in the tree. When selected it will display the advanced search options in the right panel. All searches done (via advanced or "quick" searching) will be saved as nodes under this advanced search node.

Advanced search will allow you to select which catalogs to search and you will be able to do a cone search, box search, search by name, by vla calibrator code + band + array config, and by minimum flux density + band.

Search results must show the uv min/max if present.

It was agreed that we will need a button somewhere to be able to remove a search results node to keep clutter down.

The ability to stop searches came up again, as well as the need to put some progress bar or message up so that users know something is happening.

No good way of selecting a source to use as the center of a cone search was found in the meeting so it was decided that we could put that off for now as long as it was possible to do a simple (browser) copy/paste to achieve the same thing. (It was tested and this works fine.)