ASDM Modification Planning (Post SS2)
The following summarizes our discussions on the ASDM planning
within the context of the Releases and FBTs. There are several
items that are at high priority but are prohibitive to complete
in parallel (time/resources).
I am proposing to do the measures implementation with the ASDM
which makes sense if other assumptions about priorities and
scheduling are accurate. Please review this and let me know
if this is correct or if we should change course. We're still
finishing the ACS 6.0 port but would be ready to move on this
on Nov 17th.
The implementation of measures within the ASDM will enable an
explicit representation of framed quantities; currently, all
quantities, indeed, all values have implicitly assumed units
and reference frames.
- Simplifies the handling of directions, positions,
velocities, baselines, and times in the ASDM (they will
change in type from vectors of different base types
(Angle, Double, etc) to single measure types (Position,
- Provides some integrity protection by using explicit
rather than implicit descriptions for these values;
this makes the ASDM2MS filler more robust (translation
is based on the model itself, rather than on hardwired
- Required/requested by TelCal (Robert Lucas as they
deal with measures values throughout their code).
- Implementation plan will enable an accurate schema
presentation of the ASDM; this effort has already by
prototyped by Francois Viallefond.
- Takes time (we anticipate 1 FTE month of effort = 2 months
of 'real' time).
- Precludes any other effort on the ASDM (other than SFI
and SingleDish FBT support).
Other ASDM areas that will be neglected include:
1) Real-time filler - although we don't anticipate
any technical issues with this, there would not be
adequate time before early February to complete this
work and hence would be in danger of missing the R4.1
deadline (which I understand is highly visible).
*Is this work higher priority than freezing the ASDM as
* - it will most likely require tackling 2) below.
2) Main table - the current design of the ASDM leads
to a large number of rows in the Main table; recent
experience has shown that this can cause performance
problems in both populating, storing and filling the
*Can we postpone handling this problem until later in 2007?
* - It requires effort/collaboration between:
* - Offline, Correlator, Control and TelCal
Other ASDM effort:
- CalDM representation
- The CDR4 CalDM model was implemented but appears to have
some design/technical issues; this needs resolution between
- Single Dish mapping
- Currently we are exploring IRAM 30m data to see how well the
model can represent this information.
- FV improved browser implementation
- The prototype was complete in 2006 but it requires a stable
ASDM to work reliably.
- Add some header information to the binaries for the Pointing and
ACS 6.0 Port
The ASDM plan was the following:
1) Port to ACS 6.0 using the HEAD; tag this code with something.
2) Continue working on the HEAD to accomodate some minor ASDM changes for consistency/filler ops
- We provide a list of the changes
- Heiko must be notified to change DC
- Robert must be notified to change TelCal
- tag this code with something
The hope was that step 2) could be done in essentially a day (we check-in; we test; DC adapts/TelCal adapts; they test).
3) The SFI branch is made with the correct ASDM code; we twiddle our thumbs waiting for problems.
Is this still possible to pull off?
- got okay from Jeff 20061108
ACS 6.0 port
- exception handling changes (Heiko S.)
- EntityRef changes (Michel C.)
- Tag working port as:
Minor FV changes:
configDescription:feedId -> mandatory
Feed:focusReference -> vector from vector
Scan:sourceName -> optional *not agreed*
Scan:fieldName -> vector from string
Scan:scanIntent -> vector
--> requires new signature for addRow for the Scan table
- Tag working code as:
----> ASDM will not change during SFI
----> ASDM will freeze on/about 20070101 to enable work on
Outstanding ASDM work
- CalDM inconsistency/population issues (FV/RL/MC)
- XML schema (represented in code generation)
- Main: number of rows issue
- FV browser implementation
- Pointing/TotalPowerData binary headers
- Real-time filler
- 07 Nov 2006