[evla-sw-discuss] 10/29 & 11/6 versions of EVLA e2e Software Science Requirements

[Bill Sahr]

I have finally had a chance to read the 10/29/2002
version of the EVLA e2e Software Science Requirements,
and to briefly review the 11/6/2002 version.  
My comments are concerned primarily with the issue 
of monitor and control data and refer to subsections
within the heading 2.4 Archiving.  If anything the
situation described below has gotten worse from the
10/29 version to the 11/6 version.

In general, I feel that the timescales associated with
the archiving, searching and retrieving of monitor and
control data are too late, and in some cases the assigned
priorities are much too low.  The priorities and timescales
as given in this document would force the monitor and 
control group to implement it's own archive for monitor
and control information and to maintain that archive through
Q2 2010, if not longer.  In general, I am of the opinion
that the timescales and priorities as outlined below will
definitely hinder rather than assist the development of 
the online, real-time monitor and control software.

The sections which particularly caught my attention are:

2.4.1 Archive Definition
4.1-R1.3 It is not planned to have an archive which stores
         monitor and control data until timescale C (shared-
         risk science, Q2 2007)
         This date seems _much_ too late to me.
4.1-R2 It is not planned to have search and retrieval capabilities
       until timescale C.  This timescale seems to make pointless
       the ability to store any data in the archive on timescales
       earlier than C, even though the storage of interferometer
       data and environmental data are scheduled for timescale A
       (Q2 2004).  It also seems to contradict the timescales given
       in sections 2.4.7 Archive Search Tool and 2.4.8 Archive
       Retrieval Tool which assign a timescale of A to search and
       retrieval capabilites.

2.4.4 Monitor and Control Data
4.4-R1.1  Inclusion of status of real-time system parameters
          in the archive.  pri 1, timescale C
4.4-R1.2  Inclusion of status of instrumentation (e.g., error
          flags) in the archive  pri 2, timescale D (Q2 2010)
4.4-R1.3  Inclusion of monitor point information (e.g., module
          temperatures, voltages)  pri 3 timescale D (Q2 2010)

Again, the timescales seem much too late, and the priorities
assigned to the error flags, and especially to the monitor
point information itself seem to be much too low.  If monitor
point information is to be given only a priority of 3 (desirable
but not critical) and not made available until 2010 then the
M&C group will be forced to implement its own monitor data
archive for use by engineers and technicians during development
of the EVLA.  Indeed, we already plan to do so for the test
antenna.  Do we really want to continue along separate, divorced
paths for the entire development lifetime of the instrument ?
Given the assignment of a priority of 3 to the inclusion of 
monitor point information in the archive, it is unclear that it
would be available even in 2010.  The e2e effort will certainly 
be of no assistance whatsoever to the M&C effort in the area
of monitor and control data if these priorities and timescales
are maintained as currently stated.

2.4.6-R5 The archive shall include an electronic logbook to
         record the notes of the observer and/or telescope
         operator  priority 3, timescale D (Q2 2010)
The priority is too low and the timescale is much too late.


2.4.7 Archive Search Tool
4.7-R1 There shall be an Archive Search Tool to access the Archive.
       pri 1 timescale A (Q2 2004) in 10/29 version, and
       timescale B (Q4 2005) in 11/6 version 
4.7-R1.1  The Archive Search Tool shall be operated through a 
          web-based GUI.  pri 1, timescale A in 10/29 version
          and timescale B in 11/6 version
4.7-R1.2  The Archive Search Tool shall have a CLI to interrogate the
          Archive from scripts or other Tools.  pri 1, timescale A in
          10/29 version & timescale B in 11/6 version.

The further deprecation of these items from timescale A in the 
10/29 version to timescale B in the 11/6 version seems to be a
move in the wrong direction.

2.4.8 Archive Retrieval Tool
4.8-R1  There shall be an Archive Retrieval Tool to access the Archive
        pri 1, timescale A
4.8-R2.3  environmental data is to be retrievable as 
          pri 1, timescale A
4.7-R2.5  monitor and control data is to be retrievable as
          pri 1, timescale C (Q2 2007)

There seems to be something of a contradiction between 
calling for the inclusion of monitor point information in the
archive as pri 3, timescale D, and specifying the retrieval
of monitor and control data as pri 1, timescale C.

Additionally, and finally, if we are to have archiving, searching,
and retrieval of environmental data as pri 1, timescale A, then
why not specify the same capabilities w.r.t. monitor and control
data and monitor point information also as pri 1, timescale A.
Earlier availability of this information in the archive would
greatly assist the development effort and insure the continued
integration of this information with other information in the 
archive.

Bill Sahr