[evla-sw-discuss] Software effort in the short term

[Barry Clark]

This note is a sort of sideways look at Bill Sahr's very comprehensive
note at http://www.aoc.nrao.edu/evla/techdocs/computer/workdocs/EVLA_MC.html.

I emphasize the temporal development of things, as opposed to the logical
nature of the tasks.  I am pretty much ignoring correlator issues, because
I want to concentrate on things that will come to fruition in the next
couple of years.

The software effort will fall into several phases.  Basic design 
decisions need to be made for all phases soon, but the bulk of effort
will need to be devoted to the earlier phases.  We do not have time
for the luxury of assigning people to phases to work along from start
to finish.  I am listing a reasonable phase organization.

I.  Organizational phase
    
    A.  Deliverables:
	1.  Software archive (CVS), policies for use of same
	2.  Software standards and review policy
	3.  Documentation standards

    B.  Decisions to be made
	1.  Choice of languages
	    a.  For MIB communication & computation functions
	    b.  For interface screens
		i.  Technician interface screens
		ii.  Operator interface screens
		iii.  Astronomer interface screens (Web based?)
	    c.  For observing system
	2.  Software walkthroughs?
	3.  Assignment of standards enforcement tasks

II.  Lab phase

    A.  Deliverables:  Technician interface screens for most modules

    B.  Decisions to be made
	1.  Choice of hardware for MIB
	2.  Choice of OS for MIB
	3.  Where does screen software run - MIB, technician workstation, other
	4.  How is data exchanged between MIB and technician workstation
	5.  How is data to be communicated from the observing system
	6.  How is monitor data published
	7.  How are device operational checks to be made and published
	8.  How are device excercise routines to be implemented

III.  Test antenna - single dish phase
    
The goal of this phase is that the antenna can be pointed by the observing 
system, detector data from a receiver logged, data can be merged to derive
a pointing solution.

    A. Deliverables
	1.  Command line interpreter
	2.  Timekeeping routines
	3.  Geometry routines
	4.  Antenna pointing model
	5.  Catalog and history system for antenna pointing model parameters
	6.  Monitor data archive system
	7.  Operator interface screens

    B.  Decisions to be made
	1.  Overall structure of observing system
	    a.  Division into tasks
	    b.  How tasks communicate
	    c.  System oriented around setup definition (like current VLA), or 
	       commands to MIBs, then integrate (like ALMA test interferometer)?
	2.  Language for command line interpreter - Glish?  Python?  Home brew?
	3.  CALC server, or built into observing system
	4.  Format of monitor data archive

    C.  Purchases
	1.  Machine(s) for CALC and observing system
	2.  Machine for monitor data archive

IV.  Test antenna - interferometer phase

The goal of this phase is that the antenna can be used with the rest of the
array, in a blind fashion (ie without communications with the Modcomps
running the rest of the array).

    A.  Deliverables
	1.  Translator, Modcomp input file to new system commands
	2.  Lobe rotator and phase switcher drivers
	3.  Fiberoptic IF system control, test, monitor
	4.  Data flagging system


V.  Interim operation

The goal of this phase is to support a mixed array of old and new antennas,
replacing the Modcomps.

    A.  Deliverables
	1.  Interface from observing system to CMP
	    a.  for commands
	    b.  for monitor data
	2.  Technician interface screens for old modules (L6, F4, ACU, etc)
	3.  Interface to correlator controller
	4.  Data sniffer for displays, ANTSOL, etc.
	5.  Port of current delay determination, pointing determination routines
	6.  Autophase, reference pointing

VI.  EVLA

    A.  Deliverables
	1.  Correlator software
	2.  New versions of autophase, pointing, delay solution for new
	   correlator outputs
	3.  Interferometer data archive system
	4.  Scheduling system, proposal handling system
	5.  Astronomer interface screens