Getting Started Documentation Glish Learn More Programming Contact Us
Version 1.9 Build 000 News

FAQ

Search

Home


AIPS++ Quarterly Report: 1995 Q2

Tim Cornwell, AIPS++ Project Manager
<tcornwel@nrao.edu>
July 10, 1995
Master URL: http://aips2.nrao.edu/aips++/docs/project/1995q2.htm

The state of the Project at the beginning of 1995 Q2 was close to that described by the December 1994 Review Panel report. The actions of this last quarter were targeted towards many problems in the Project, but mainly those identified by the Review. The following major changes occurred:

  1. A full-time Project Manager was appointed to provide leadership and direction,
  2. The strategy to be followed by the project has been defined: we will test and extend the infrastructure library by writing a few key astronomical applications,
  3. We have instituted development plans, schedules, milestones and resource allocations; all of which are required to run a project of the size of AIPS++,
  4. We have instituted closer ties with the astronomical community within NRAO.

Overview of developments in 1995 Q2

Inside NRAO, AIPS++ is now treated as a construction project with dedicated staff and budget. The AIPS++ group splits into two principal groups: one in Charlottesville concerned mainly with support of Single Dish processing, and another in Socorro, concerned with Project Management and Synthesis support. In addition, NRAO plans dedicated AIPS++ programmers at both the Green Bank and Tucson sites.

The immediate goals of the Project have been defined to be consolidation and testing of the AIPS++ Library, and development of a few key applications chosen to provide unique astronomical capabilities. The long-term goal of the Project has been defined to be the achievement of functional equivalence to AIPS by 2000. At that point, AIPS will be a very small subset of AIPS++ and most applications areas will look quite different from the corresponding areas in AIPS.

A development plan for the next 12-18 months has been instituted to provide a coherent overall picture of the direction of the Project in the intermediate term. Tracking of progress in AIPS++ is now performed using a Target Dates mechanism.

The following applications are now present and being developed further:

Documentation for both Users and Programmers is now available via the World Wide Web (WWW) and is being developed further.

The following Infrastructure Library changes have occurred:

Intellectual developments are also crucial for the long-term success of AIPS++:

Management

The structure of Project management is now simpler and leaner. The structure above the Project Manager is easier to deal with than the previous Steering Committee and affords the correct degree of latitude to act. There is currently little substructure within the Project. With a couple of exceptions, either I or my deputy, Brian Glendenning, intend to be very closely involved in overseeing all areas of the Project. The main exceptions are the Single Dish development, for which Bob Garwood is responsible, and the Coordinate Classes development, for which ATNF, in the guise of Neil Killeen and Wim Brouw, is responsible.

The overall strategy and the plans for the next 12-18 months are described in the AIPS++ Development Plan.

We have adopted a more formal way of structuring the work that is to be done. First a specification is written, usually by a scientist. Secondly, a design document is prepared by a programmer/architect. Third, a development plan is put together to state what is done when and what the required resources are. Then implementation of the design is performed. We then expect testing and validation to be performed. Tracking of progress towards these subtasks is via a list of target dates that is updated weekly (see an example in Appendix B). We plan to extend slightly this formality of work structuring by adopting some ideas developed by the NFRA Software Engineering Group.

Work on the critical path is limited to those participants in AIPS++ who have demonstrated a capability to devote a known and predictable fraction of time to AIPS++ work. This includes all those in the NRAO AIPS++ Project division, some NRAO programmers and scientists outside the division, and a (small) number of people at consortium sites. The intention of this restriction is to achieve reliability of prediction of milestones, something that is vital for efficient planning. Participation by others is eagerly welcomed but is for the moment restricted to tasks not on the critical path. An example of the latter is the possible development of additional, supplemental User Interfaces.

Links to Astronomers

We have started to strengthen the links of AIPS++ to the astronomical community by involving specific NRAO astronomers in a number of aspects of AIPS++. Astronomers are involved in writing specifications of elements of AIPS++ (e.g. the properties of the AIPS++ tasks SDCalc and SDImager that are to be used for analysis of singled dish spectral data and single dish data for imaging). An astronomer (Alan Bridle) is developing the User Documentation. We have chosen to try to form astronomer-programmer alliances of the sort described by the Review Panel. At the moment, this is limited due to the overall paucity of applications. However, we expect the situation to improve in the next two quarters.

In Spring 1996, we intend to start more formal interactions, via an AIPS++ Scientific Advisory Group.

Consortium Interaction

Distributed development of AIPS++ has not yet been successful. It only seems to work in those few cases where work of a specific and limited nature is to be carried. Examples of successful distributed development include the AIPS++ Table system (a collaboration between Brian Glendenning of NRAO and Ger van Diepen of NFRA), the Code Management System (developed mainly by Mark Calabretta of ATNF), and the AIPSView visualization task (developed by a group at NCSA). Failures of distributed development are the Green Bank model for calibration and imaging, and the coordinate system (until the recent work by Brouw). The message is clear: development of complex things requires either one person working mainly alone, or good communication between a number of collaborators. Bearing this in mind, we have deferred some of the more complex development until the right people can be in the same place (e.g. Mark Wieringa, Brian Glendenning, and Tim Cornwell for Synthesis Calibration and Imaging, which will occur in September 1995 when Mark visits the AOC for an extended period).

Relations amongst the principal AIPS++ consortium members have been fruitful and basically harmonious. The main interaction between the Project Management (Tim Cornwell) and the principal consortium site managers (Jan Noordam, Dick Crutcher, and Neil Killeen) is via e-mail and telephone. So far, this seems to have worked reasonably well though there have been a few examples of poor or non-existent communication (mainly due to the Project Management).

The core consortium members, ATNF, BIMA, NFRA and NRAO, are contributing significant personnel and have, with only a few passing exceptions, produced what has been expected of them by the Project Management. The main contributions of the various consortium partners have been:

Plans for 1995 Q3

The development plans for the next 12-18 months are given in the AIPS++ Development Plan (see Appendix A). Here I summarize the expected developments over the next quarter. These can be assigned to various categories:

In Single Dish support, we will support the GBT systems integration work on the NRAO 140' telescope. This requires reading of astronomical and engineering data into AIPS++ Tables for manipulation and plotting using the Glish User Interface. We will continue to support the On-The-Fly mapping tool developed for and installed at the NRAO 12m telescope. We will specify SDCalc (for analysis of Single Dish spectra) and SDImager (for imaging from Single Dish data). We will choose one of these for high priority development by the Single Dish group.

In Synthesis support, we will develop tools for advanced imaging using the Briggs NNLS deconvolution algorithm in a selfcalibration cycle. This will be important for high dynamic range VLBI imaging. We will develop a complete calibration package for the ATCA. With the recent development of a very general formalism for UV calibration, it may be possible to make this largely consistent with calibration for other telescopes such as WSRT, VLBA and BIMA. Otherwise, calibration of these telescopes is expected to await completion of the ATCA calibration, due in early 1996.

In AIPS++ Infrastructure, we will develop the Glish system (the User Interface and Control Hub) according to the Glish Development Plan. We will complete the design of and start implementation of the support for task control within AIPS++. We will complete design of the Coordinate classes and start upon implementation. We will add support for hyper-tiled access to the AIPS++ Table System.

In Visualization, we will continue to add functionality to AIPSView as needed by other AIPS++ applications (for example support of selection of regions of interest and visualization of visibility data). In addition, extension of AIPSView to include 3D visualization will continue. We will select a package for line graphics inside AIPS++.

In the System area, we will add support for automated testing of AIPS++ classes. We will investigate use of the new Sun native code C++ compiler. We will also re-assess any possible use of the IBM compiler.

In Personnel, we will complete hiring of two positions in Socorro, one of which will be supported initially by funding from NFRA. Wieringa will visit Socorro for a 6 month period starting in September (Funding for this comes from ATNF and NFRA).

In Documentation, we will continue the development of the overall documentation system, paying particular attention to the programmer level documentation. We will remove defunct design documents and update those that apply to current code but are out of date.

In Management, communication between consortium partners will be improved by frequent visits of Project Management to consortium sites. We will request new requirements documents from consortium partners to allow planning to target the highest priority areas of development. We will continue to formalize work structuring by adopting approaches similar to those proposed by the NFRA Software Engineering Group. A number of long term, operational issues must begin to be addressed. For example, the Project must decide how system versioning and updating are to be accomplished. Code acceptance procedures must be made realistic and enforceable.

Areas of Risk

To summarize the progress that has occurred in the Project, I think that it is illuminating to consider explicitly the risks inherent in the Project. My current list of risks is given below. The current answers are given after each point.

Requests for action or approval

  1. I expect the Executive Committee to review this document closely and to comment as it pleases. However, I would specifically appreciate comments on the Areas of Risk. Are all the areas of risk identified? For those given, are the solutions plausible?
  2. I am not in favor of starting another active AIPS++ site until the current sites can be shown to collaborate well. An important test of a higher level of collaboration is the development of the Coordinate Classes that ATNF has agreed to pursue. This will be done away from the AIPS++ Center but will make significant requests of the Center to address any required changes in the Infrastructure Library. In addition, personnel will be allocated by the Center to work on the Coordinate Classes under the direction of ATNF managers (Killeen and Brouw). If this is successful, we will try further collaborations of this sort. For work that cannot be split up in this way, it is important that the Project be given more manpower at the Center. With this in mind, I request that consortium members consider the possibility of sending people to work at the center for extended periods (6-12 months).
  3. I strongly prefer to make publicly available many of the Project management documents. With this in mind, I would like to make this report publicly available, preferably via the WWW.

Appendix A: AIPS++ Development Plan

The development plan is updated every few months. Here I attach the version from 1995 June 9. The current version is available on the WWW.

AIPS++ Development Plan

Tim Cornwell AIPS++ Project Manager <tcornwel@nrao.edu>
Last Modified: 9 June 1995
http://aips2.nrao.edu/aips++/docs/project/devplan.html

Introduction

This document describes and records the development plan for AIPS++ from March 1995 (when new management of AIPS++ was instituted) until mid-1996. The plan will evolve and change with time. Revisions since the last version are shown in bold. Items subject to continuing discussion are given in italics. With the exception of this section, which has some general observations, the style of this document is direct. It says what we will do with relatively little emphasis on saying why we will do it or who will do it. This plan is mirrored in the AIPS++ Target Dates where names and time-scales are given. I have matched the goals here to the personnel available rather than give a wish-list. I have also described areas where we will not be active.

First I give some general observations that drive the plan:

AIPS++ is conceptually broken down into various sub-components: Infrastructure, System, Documentation, Single Dish Processing,Synthesis Processing, Image Analysis,Visualization, Observing support, and Management.

Infrastructure

System

Documentation

Single Dish Processing

Synthesis Processing

Image Analysis

Visualization

Observing support

Management

aips2-workers at the beginning of every week.
tcornwel@nrao.edu

Appendix B: AIPS++ Target Dates for June 26, 1995

The Target Dates are accessible via the WWW. Here I attach the list of Target Dates from 1995 June 26. These are discussed and updated during the Monday Morning Meeting held every week. In addition, each active worker is expected to submit a progress report at the end of every week.

Date               Target             Who?                Original           

6/17/95            Specification of   Noordam             9/1/95             
                   WSRT calibration                                          

6/26/95            Quarterly Report   Cornwell            6/26/95            
                   to AIPS++ Board                                           

7/1/95             Specification of   Liszt               7/1/95             
                   SDCalc                                                    

7/1/95             Image masks and    Roberts             7/1/95             
                   weights                                                   

7/1/95             Plan for           Cornwell            7/1/95             
                   programmer                                                
                   documentation                                             

7/1/95             Specification of   Cornwell/Emerson    7/1/95             
                   SDIMager                                                  

7/1/95             Cleanup of         Cornwell            7/1/95             
                   defunct design                                            
                   docs                                                      

7/5/95             Design of generic  Olnon               7/5/95             
                   filler                                                    

7/5/95             Plan for work in   Cornwell            6/1/95             
                   Synthesis                                                 

7/7/95             Design of          van Diepen          9/1/95             
                   Hypercube/tiled                                           
                   storage manager                                           

7/10/95            Design of design   Cornwell            6/15/95            
                   documentation                                             

7/15/95            GBT initial test   Garwood             7/15/95            

7/17/95            Design of          Glendenning         7/1/95             
                   Tasking, etc.                                             

8/1/95             Second Hire in     Cornwell            11/1/95            
                   Socorro                                                   

8/1/95             First Hire in      Cornwell            8/1/95             
                   Socorro                                                   

8/1/95             Specification of   Wieringa            8/1/95             
                   ATCA calibration                                          
                   package                                                   

8/1/95             Implement          Wieringa            8/1/95             
                   polarization                                              
                   solve and apply                                           

8/1/95             Design of          Young               8/1/95             
                   Programmer                                                
                   Documentation                                             

8/1/95             Select Graphics    Shannon             8/1/95             
                   package                                                   

8/1/95             Automated testing  Calabretta          7/1/95             

8/1/95             Move Functionals,  Glendenning         8/1/95             
                   Fitting                                                   

8/1/95             Move Lattices,     Roberts             8/1/95             
                   Images                                                    

8/1/95             Plan for First     Cornwell            8/1/95             
                   Generation Image                                          
                   Analysis                                                  

8/15/95            Design of          Brouw               8/15/95            
                   Coordinate                                                
                   classes                                                   

9/1/95             Development plan   Brouw               9/1/95             
                   for Coordinate                                            
                   classes                                                   

9/1/95             Move               Garwood             9/1/95             
                   MeasurementSet                                            

9/1/95             Move Correctors    Glendenning         9/1/95             

9/1/95             Implementation of  Glendenning         9/1/95             
                   Tasking design                                            

9/15/95            12-m Continuum     Garwood             9/15/95            
                   OTF data handled                                          
                   by OTF tool                                               

10/1/95            Select GUI         Shannon             10/1/95            

10/15/95           GBT second phase   Garwood             10/15/95           
                   test                                                      

10/22/95           ADASS meeting      Cornwell            10/22/95           

12/1/95            Reimplement FITS   Young               12/1/95            
                   classes                                                   

1/1/96             Plan for Image     Cornwell            1/1/96             
                   Analysis                                                  

1/1/96             Specification of   Teuben              1/1/96             
                   BIMA calibration                                          

2/1/96             Design of          Garwood             2/1/96             
                   SDImager                                                  

2/1/96             Specification of   Beasley             2/1/96             
                   VLBA calibration                                          

3/15/96            GBT third phase    Garwood             3/15/96            
                   test                                                      

5/1/96             Implement alpha    Garwood             5/1/96             
                   SDImager                                                  

7/1/96             Design of SDCalc   Garwood             7/1/96             

10/1/96            Implement alpha    Garwood             10/1/96            
                   SDCalc                                                    



Appendix C: AIPSView Capabilities

AIPSView is now in a second alpha test. It has the following capabilities.


Appendix D: Personnel

The following changes have occurred:

The total roster of programmers and managers working on AIPS++ is shown below, along with percentage time allocated to AIPS++ work:


You could now go back to the:


Copyright © 1995,1996,1999,2000 Associated Universities Inc., Washington, D.C.

tcornwel@nrao.edu

Modified: 2003/07/16 04:06:31 GMT
Installed: 2003/08/21 11:37:44 GMT