 News |
FAQ
|
| Search
|
Home
|
| Getting Started | Documentation | Glish | Learn More | Programming | Contact Us |
|
| Version 1.9 Build 1367 |
|
T.J. Cornwell, NRAO
1998 October 16
The Third Beta Release was made on October 15. Another significant project milestone is the first Nature paper to have used AIPS++: the August 13 edition has a Parkes Multibeam paper on a new leading arm for the LMC (Putman et al.).
In Single Dish support, we have made a number of enhancements and bug fixes to the dish environment. The full SDFITS convention is now understood by dish. There is full support for multi-polarization data in all dish operations. The command line interface to dish was significantly improved and the internals of dish were somewhat reorganized. A fair amount of real single dish data from a number of different telescopes has been collected for use in demonstrations and testing of dish.
In support of the GBT, we have developed a basic continuum commissioning software suite in AIPS++. A script was written for reducing total power scans from the GBT spectral processor. These reduced data can then be processed using dish. The AIPS++ RFI monitoring program for use in Green Bank was revised based on feedback from the RFI committee in Green Bank. This program is now in use by the 140ft/GBT operators.
Single Dish work suffered in Q3 from the departure of Jeff Uphoff from NRAO. Joe McMullin moved from Green Bank to Charlottesville. Both Bob Garwood and Joe McMullin are involved in supporting the GBT through regular 2 to 3 day visits by at least one of them to Green Bank each week. A replacement for Jeff Uphoff, to be based in Green Bank, has been advertised.
In Synthesis support, work has followed the development priorities and guidelines as stated in the report for the immediately preceding quarter. This includes development on a broad front in key synthesis areas such as imaging and calibration, in line with the long-standing AIPS++ policy in this area, but with some resources devoted to an additional development stream for a limited end-to-end data reduction capability ("thin-path" capability) for instruments such as the VLA. In addition, synthesis development for instruments where AIPS++ is in the critical path (such as WSRT), remains a priority. The overall aim of the current development is to expand the scientific capabilities and user base of the system as a strong priority. The current emphasis is on application development over library development.
Consultations have continued, both internally and externally, on the approaches to calibration and imaging being adopted in synthesis development within the project. This is in keeping with the requirement that the system serve as a data reduction environment for a broad range of instruments, and is consistent with the distributed expertise that exists in the project in this area. Travel by A. Kemball to NFRA in June and by W. Brouw to the AOC in July were helpful in this respect.
The proposed data format revision to V2.0, finalized in previous quarters has seen little change in this quarter, with the exception of some changes for specialized sub-tables (A. Kemball). A plan has been adopted for the implementation of the revised format immediately following the third beta release. The revised format is required for VLBI and other more specialized synthesis development, and will be be implemented by M. Wieringa.
Further WSRT specific utilities for visibility data display and manipulation have been developed by J. Noordam and others at NFRA in support of the WSRT upgrade and TMS commissioning. These utilities are primarily Glish-based, but include other developments in C++, including a holography application. Further details can be found in the NFRA contribution to the report. An additional WSRT synthesis development has involved participation in the evaluation of the PIM ionosphere model, which is being reviewed elsewhere at NFRA. The code and data formats required to support this model in AIPS++ have been considered.
Work on data fillers has continued this quarter, with an update to the BIMA filler (D. Roberts and P. Teuben), and initial design and development work for a prototype VLA filler (R. Marson and A. Kemball), in keeping with the "thin-path" strategy for the VLA. Basic components for graphical display and editing of visibility data are required for "thin-path" capabilities for all instruments, and are an important element in developing early automated data reduction capabilities. This is a priority at several instruments. Design work for uv display components has proceeded this quarter (A. Kemball), in consultation with the developers of the Display Library (D. Barnes and N. Killeen). This work will continue in collaboration with P. Barnes (NRAO), who joined the project in June. The Glish visibility display tool "visplot" has also been revised and extended during this quarter (T. Cornwell).
A significant focus has been retained in this quarter on the development of cross-calibration capabilities across a broad front to expand the current calibration facilities in the system (A. Kemball). A formalization of calibration table formats for visibility- and image-plane effects has been completed, together with the framework for the continuing development of more advanced calibration parametrization, interpolation and solvers.
A primary development effort in imaging has concerned mosaicing, which has seen renewed emphasis in this quarter (M. Holdaway and T. Cornwell). A mosaicing algorithm development workshop is planned for July 1999 by M. Holdaway, who joined the project in July, Michael Rupen and Tim Cornwell. Mark is taking the lead in mosaicing implementation in AIPS++. This has proceeded as outlined in the development plan for mosaicing, with initial emphasis on expanding the primary beam correction software and related matters. T. Cornwell has worked on wide-field imaging questions relevant to space-based arrays, as part of a collaboration with NRL. As part of this effort, the uv-data simulator has been updated (T. Cornwell). In addition, continued maintenance and revision of the imaging software has continued throughout the quarter (T. Cornwell), including updates to the source component model infrastructure (R. Marson). It is anticipated that the lead on wide-field imaging will be taken over by T. Cornwell, with the tragic death of D. Briggs in July.
Finally, the application simpleimage has been improved in a number of ways: it can process multi-source and multi-spectral-window data, and the Glish commands used are displayed to the user and also written to a file for use as an template scripts.
Vacant positions are currently advertised for parallel synthesis application development, and wide-field imaging development, both to be filled at the AOC.
In Glish, preparations began for the upcoming beta release of AIPS++ and public release 2.7 of Glish. The user manual was updated to document all of the additions since the last public release of Glish. The whole manual was reviewed in the process of doing this update.
In addition to updating the manual, a few new features were added:
Finally, a number of bugs were fixed:
In Measures, Wim Brouw accomplished the following:
Not much work was done on the planned extensions for this quarter, with the exception of the unification of the coordinate system tranforms. Work on Earth tides and comet-like ephemerides has not progressed.
In AIPS++ Infrastructure, little new work was performed apart from bug fixes and cleaning up:
In Image Analysis, all work was connected with regions in images. See Neil Killeen's report for details of what he did directly in Image Analysis in this quarter. In addition, Ger van Diepen has worked on the region classes (see NFRA report for his details). As a result, we now have:
In summary, we pretty much achieved what we planned for this quarter. A small number of outstanding items have been moved to next quarter's plan. None of the above changes made their way into the system as we were unable to test well enough before the beta release. They will be checked in very soon after the beta release.
In Display work, David Barnes acheived the following:
In the System area, we have continued to suffer the loss of Jeff Uphoff, who was dedicated 50% to system work, and the move of Joe McMullin to Charlottesville. Logistics demand a full-time astronomer at the GBT so we will end up down half a person for system work. Some small relief comes from the fact that Ger van Diepen has agreed that starting September 1st he would take over the role of technical manager from Brian Glendenning. In September he visited Socorro to discuss the tasks and the system work with the AIPS++ management. It resulted in a system plan which is laid down in note 220. It describes the future plans in the system area. Ger will work with Darrell Schiebel on implementing a triage approach to our systems work until a more acceptable long-term solution is found.
Wim Brouw's data proposal has been accepted and will be implemented in the next quarter. It results in a clear separation of the source code and the data files needed to run AIPS++. A rearrangement of the code tree is deferred until next year.
It appeared that gcc 2.7.2 and Solaris 2.6 do work together very well. It is investigated if the available patches solve this problem.
Building the documentation was more reliable than in previous quarters, but there are still a few small problems with latex2html. It is being investigated if other tools could be useful to build the documentation.
The GNATS system was found to be somewhat unfriendly and cumbersome to use. It is being investigated if another bug reporting system (e.g. the one used in the GNOME project) can be used instead.
Brian's last deed was the preparation of the beta release, which was ready on October 15th. He documented the process to make the creation of future beta releases easier. The beta release is based on gcc 2.7.2, but it appeared that for Linux Red Hat 5.1 the egcs compiler had to be used.
To provide more rapid feedback to our testers and users, it is planned that from November on monthly binary releases will be generated based on the most recent stable inhale. A few scripts will be developed to create and install such releases.
Shortly after the beta release a few changes will be done to make our use of C++ more compliant with the standard:
The current project compiler is gcc 2.7.2. In the coming quarter it will be investigated if other compilers, notably egcs, run well on all supported platforms. If so, it will be decided if another compiler will be the project compiler.
Static variables are sometimes causing problems when using exceptions or shared libraries. Some efforts will be done to reduce the number of statics and to create and destruct them in a way that is harmless for exceptions and shared libraries.
In the area of Parallelization, in the third quarter of 1998, the parallel group suffered the loss of Dan Briggs, who was a postdoc working on parallelization and had just started the parallelization of a wide field imaging application within AIPS++. Dan was killed in a skydiving accident. This tragedy has obviously affected our small parallel group. In addition, our group leader, Brian Glendenning, has moved on to a position in the Millimeter Array (MMA) and Athol Kemball has taken that role in his stead. The changes in the group are reflected in the specific focus on a limited number of parallel efforts.
The most significant accomplishment has been the development of parallel algorithm iterator by Wes Young. This algorithm iterator allows an application programmer to implement high level algorithms which loop over independent pieces of data. The obvious test case for this is independent spectral line channels. Wes has implemented a Clark CLEAN algorithm using these classes.
In preparation for an NSF site review at NCSA, Doug Roberts benchmarked the parallel Clark CLEAN application. This test was in preparation for a large scale imaging and deconvolution project that the parallel group is enabling. This significant project entails making maps (> 6K on a side) and a large number of spectral line channel (>100). The parallel group is helping Dave Westphal to process this large HI data set of M33. In initial tests performed for the NSF site review, good speed-up (almost a linear slope of 1.0) was observed between 1-16 processors. However the slope was substantially worse (slop of 0.5) between 16-32 processors. The group is investigating the performance. It is possible that this has exposed an I/O bottleneck. This test suggests that understanding I/O performance is important to ultimate parallel performance.
Roberts has been using the SGI installation to process and analyze data. Most tasks appear to work in the SGI installation. Until late in the quarter a problem existed with the sky (now imager) application, which was exposed due to the big image (2K on as side) work exercised by the parallel group. This problem has been fixed it is likely that soon there will be a blessed SGI release.
We have been in contact with several National Computational Science Alliance (Alliance) partners. We have had meetings with the Parallel I/O group headed by Dan Reed, UIUC computer science department chair. Their group has decided to install AIPS++ on a linux system and will replace the low-level I/O calls with calls to an instrumented library that they have developed. They have the installation complete now and are now beginning to instrument our code. After this is complete (last quarter of the year), we should be able to investigate I/O usage and identify bottlenecks (similar to the one that might be responsible for the slower speedup after 16 processors mentioned above). This will also be useful in investigating the use of a parallel I/O library that is available as part of the MPI-2 standard. The MPI-I/O is available now, although other aspects of the MPI-2 standard are under development.
Another possible Alliance partnership is with the Globus group. Globus is a project to design and build middleware for authenticated data transfer and control across the computation grid. The Globus group is lead by Ian Foster from Argonne National Laboratory. At NCSA, a group involved with real time data transfer is investigating using Globus. We are in contact with that group and may use their experiences to guide the potential uses of Globus in our parallel execution environment for AIPS++ on parallel (or clusters of parallel) computers.
Finally, we have nearly completed the installation of a NRAO frame-relay network node at NCSA. This will allow a guaranteed bandwidth for Wes Young to continue development in Socorro on NCSA machines, such as the Origin 2000. In addition, astronomers at NRAO should be able to transfer relatively large data sets to the NCSA computers via overnight ftp sessions and to use visualization and graphical user interfaces across the higher-speed intranet connection.
In Documentation, we (finally) produced a document to help new users: Getting Started in AIPS++. This is a large expansion by Anantha of a seed document written by Tim Cornwel some time ago. It explains the weird and wacky world of AIPS++ in terms that astronomers (mainly Anantha, Michael Rupen and Bob Hjellming) are comfortable with. This document is a key ingredient of the third beta test.
We have proceeded with incremental improvements to the Web Site. For example, Glish and the Quality Assurance Group both now has their own pages, and we have a page devoted to job opportunities within AIPS++. Also we now have a simplified Web address: http://aips2.nrao.edu/ (Unfortunately, AIPS++ cannot be used since most browsers will not accept it).
In Management, the replacement of Brian Glendenning by Athol Kemball (as Deputy Project Manager) and Ger van Diepen (as Technical Leader) has proceeded extremely smoothly. This was aided by a very successful visit of Ger to Socorro in September, and by weekly phone conferences between Athol, Ger and Tim Cornwell.
We have started monthly meetings of the consortium site managers. This consists of Tim Cornwell, Athol Kemball, Ger van Diepen, Jan Noordam, Doug Roberts, and Neil Killeen. With Neil's good grace we have these meetings at 11pm Sydney time (13UT) and so all are able to attend. We find this forum very useful for discussing issues such as personnel, budgets, planning, that are not suitable for discussion in the long-standing Monday Meetings of all AIPS++ Project members.
The AIPS++ Project and the MMA Computing division are jointly sponsoring a meeting on Mosaicing Techniques to be held next July. Mark Holdaway, Michael Rupen and Tim Cornwell are organizing this meeting, at which it is hoped that the state-of-the-art in mosaicing will be discussed and documented by those most active in this area.
JCMT has approached AIPS++ with the desire to use AIPS++ as the platform for the new ACSIS multi-beam/correlator project. We have agreed that we expect that this will be possible, pending negotiations on the exact requirements. We expect that the support required from the Project will be small. We plan to discuss more details in a visit by Tony Willis and others to Socorro in mid November.
In the Quality Assurance Group, work was performed on:
The pre-release code was frozen on October 2, and was tested, principally by testers and project staff at the AOC, and released on October 15. Next, we plan to move to a different scheme for distributing the system. Instead of a large beta release every 6-12 months, we plan to release binaries for the stable builds. Every month, we attempt to certify one build as being stable and complete. It has to pass a number of tests, including both our standard unit-testing and interactive, Glish-based tests. It is then placed on our ftp server in order that those wishing to acquire a good-quality build can do so using our inhale mechanism. In the new approach to distributions, we will build binaries for each stable version, and also make these available on the ftp server. Installation of a new stable system is then possible approximately every month. Ths has the advantage for the users that new features and bug fixes appear rapidly, and for the project, that we can more easily and quickly get feedback on new capabilities. It is envisaged that this new ``trickle-out'' approach will continue until the first public release next year.
The response of our established beta-testers to this new approach has been positive (or at least neutral!). From a poll, it seems that building on Solaris and Linux RedHat 5.1 will be sufficient, and so the builds can be performed at NRAO.
Compared to the second beta, the third contains substantial changes and additions in essentially all areas:
In Single dish support, we expect to receive a significant amount of user feedback from dish as a result of its inclusion in the upcoming AIPS++ beta release. This is the first inclusion of dish in an AIPS++ beta release. In addition to responding to user feedback, we plan on adding new operations to dish and to improve aspects of the user interaction with dish. Dish will be able to function fully from the command line and the user will be able to start and stop the dish graphical user interface without exiting AIPS++. The dish state saving mechanism will be more flexible to allow for the state to be saved and restored while running the program and to allow the user to have multiple states stored on disk. We expect to start to integrate dish with the GBT commissioning tools. The command line interface to the dish plotter will be cleaned up. The conversion between an AIPS++ MeasurementSet and an SDFITS convention binary FITS file will be enhanced to encompass the full SDFITS convention. dish will then be changed to interact directly with an AIPS++ MeasurementSet. The existing GBT fillers will be reworked to improved their speed and usefulness with on-line data. We will begin to work on the pointing analysis and modeling of the telescope structure for the GBT commissioning support. Work is expected to start late in the quarter on single dish imaging using the same imaging scheme as used in the synthesis imaging already in place in AIPS++.
In Synthesis, we expect to respond to beta-testers, fix bugs, but also continue on our planned development plan: adding more functionality in filling of data, calibration, and imaging. In particular, the mosaicing capabilities will be augmented by the addition of new classes to support more complete models of primary beam effects, such as beam squint for the VLA. This example will give us entirely new capabilities for polarimetric imaging of large fields with good accuracy, as is important for the Sun, for example.
In Measures, we plan to finish the coordinate system (directions, baselines, uvw etc) unification and if possible do some work on Earth tides.
In Images, we plan the following:
In Display, we plan the following:
In Management, we will start planning the next meeting of advisory groups. Our tentative plan is to split the STAG into two groups, one of users and one of technically competent experts, as we always have envisaged being the ultimate situation.
We will make presentations at ADASS in November, including a demo of the third beta.
In mid November we will hold a meeting of those project members working on applications. The purpose of this meeting is to come up with styles and standards for writing applications in AIPS++. Those expected to attend are: Cornwell, Kemball, Garwood, Schiebel, Killeen, Barnes, van Diepen and Noordam.
In the Quality Assurance Group, we will:
This quarter the ATNF had 4 people working in AIPS++. These are Neil Killeen (25-75%, also local manager), Wim Brouw (75%), Mark Wieringa (30%) and David Barnes (75%).
In the next quarter, Wim's availability will be markedly decreased as he he is temporarily (about 3 months) seconded to the ATOMS project (our on-line OO software project). Wim will work in AIPS++ at about the 20% level next quarter.
Our systems (Solaris/Gnu and Digital Unix/Gnu) were generally stable this quarter. The documentation pain of last quarter has eased somewhat, but the need for too frequent software (such as LaTeX) upgrades persists.
David visited NCSA for two weeks in order to maximise information transfer from the departing John Pixton and to try and get as much of the DL infrastructure under control as possible before John left; this was a successfull visit as we believe the loose ends are few (and known).
Ger van Diepen visited Epping for two weeks. This time was spent mainly on discussions and code implementation revolving around image regions. Ger also spent time with David learning about the Display Library.
Overlapping with Ger, Tom Oosterloo was also here. Apart from joining us in the regions battles, Tom worked with us generating some design ideas on how the basic core DL applications might look to the user.
These periodic visits that we do are always incredibly successful. Our progress accelerates enormously during them.
Commissioning of narrow band multibeam systems is beginning. The multibeam software has now been checked into the AIPS++ system and the ATNF will play a larger role in the care of multibeam software. Previously it was created and looked after by the multibeam science teams (some of whom were ATNF staff). This change comes about as the multibeam receiver moves more into a National Facility role. The principle ATNF people now involved in the multibeam software are Lister Staveley-Smith and Mark Calabretta.
Wim has spent much of his time this quarter on updating and related work on Measures, as discussed above.
Other activities:
David Barnes was totally responsible for all changes to the Display Library discussed above.
Neil writes:
Virtually all my efforts this quarter went into image regions related issues. I have learned this quarter (again) that having a system with enormous scope, power and flexibility, does not always make it easy to deliver. One's eyes light up with the excitement of the boundless possibilities, and the trick of limiting those possibilities is non-trivial !
Neil's time went on:
Mark has again had little time this quarter for AIPS++. Partly this was due to the heavy floods at Narrabri, holidays and other local issues. It is expected that next quarter Mark's AIPS++ time will increase again.
He has spent his AIPS++ time on:
During the past quarter we had 3 people (1.3 FTE) working on AIPS++: John Pixton (20%), Harold Ravlin (10%), and Doug Roberts (100%). In addition, NCSA is funding the salary of Wes Young at NRAO-Socorro and will shortly fund the salary of a second NRAO person. Accomplishments during the past quarter are described above under Parallelization.
Local project members: Ger van Diepen (GVD), Michael Haller (MXH), Jan Noordam (JEN, local manager).
Michael Haller has been transferred to the TMS project for the remainder of 1998. He will continue in his role as local AIPS++ system manager, and will participate in the QAG in the future.
Bob Campbell (postdoc at NFRA) has made a large contribution to the definition of ionospheric calibration in AIPS++. He will remain involved in the near future.
This quarter, the NFRA contribution in FTE's was as follows:
| Global AIPS++ | Local AIPS++ | Other | |
| Ger van Diepen: | 60% | 20% | 20% |
| Michael Haller: | 10% | 30% | 60% |
| Jan Noordam: | 30% | 40% | 30% |
The new WSRT on-line system (TMS) continues to rely heavily on AIPS++ modules, particularly Tables and Measures. Is has now also adopted Glish. With the exception of the latter, which is still a bit fragile at times, these various modules have performed with very few problems.
The first generation of local AIPS++ applications (in Glish) for off-line uv-data inspection and the extraction of set-up parameters works satisfactorily. In order to deal with the full range of tools that are required, a somewhat more able uv-data structure (the 4D 'uvbrick') has been developed, together with its user interface ('msbrick'). It is based on the 4D data structure that is extracted from the MS by the ms DO. Its methods include a limited form of selfcal that solves for all kinds of instrumental parameters using a bright central point source calibrator (which may be polarised). In addition, astronomers may manipulate and inspect uv-data in various ways, using basic operations on uvbricks. Altogether, it amounts to a 'mini-package that fills a niche that is complementary to 'official' AIPS++, somewhat like the single-dish 'Dish' application. It is fully generic in the sense that ifr-selection is no longer WSRT-specific.
The 'HOLOG' DO has been developed to reduce holographic observations. For the time being it is a bit WSRT specific, but without too much work it should be possible to turn it into a general tool.
The program ms2scn (which convert a MeasurementSet to a Newstar SCN file) has been enhanced to deal with multi-field, multi-spectral window observations. It also uses the ANTENNA subtable to find the telescope number for an antenna-id.
UVFITSWriter writer has been extended with writing the TY and GC tables (system temperature and gain curve) to be able to load these data in classic AIPS.
Harro Verkouter (JIVE) has continues his work on a more permanent version of the UVFITS writer. He has finished the multi-source part, and it has been agreed that he will take care of multiple spectral windows.
Access to tables has been optimized by Ger van Diepen for accessing parts of a column. The class RefRows has been developed for this purpose to have a uniform interface to the data managers.
The TiledShapeStMan has been developed. It resembles TiledDataStMan, but chooses the hypercube automatically based on the shape of the data array. It frees the user of the burden to add or extend hypercubes, but offers less flexibility than TiledDataStMan.
In the next quarter the ability to store records in a column will be added. It is needed to be able to store regions and parameters in a table.
An introductory article about tables has been written for the AIPS++ newsletter. In the next quarter some more articles will be written discussing how to use tables and the query language. Also a note will be written explaining the tiled storage managers.
PagedImage has been changed such that it is possible to access a readonly image. For this purposes LogTable needed to be changed as well to access readonly log tables. The glish interface to log tables has been changed similarly.
Ger van Diepen worked with Neil Killeen on regions in pixel and world coordinates. The LCRegion classes make it possible to define a region in pixel coordinates and are used to access the required pixels from the lattice/image. The basic LCRegion classes are LCBox, LCEllipsoid, and LCPolygon. Compound LCRegion classes make it possible to define the union, intersection, difference, complement, and concatenation of LCRegion's. LCExtension is a specific compound class to extend a region along straight lines in other dimension(s). The LCSlicer class is similar to LCBox with the addition of strides. LCSlicer is not an LCRegion, because it cannot be used in compounds. The WCRegion classes make it possible to define a region in world coordinates. Similarly to the LC classes there are compound classes to make a union, etc. of WCRegion's. When a WCRegion is used for an image, it is converted to the appropriate LCRegion using the coordinate and shape information of the image. Since it knows the world axes of the region and the image, it can adjust the order of the region axes to match the image axes. When the image has more axes than the region, the region will be extended automatically.
The regions are almost finished now and will be checked in after the beta release. The classes SubLattice and SubImage will be able to handle them. The next step is changing the lattice expressions classes such that they handle a possible mask associated with a region. The LEL classes have been changed, but they need extensive testing before they can be checked in. This will be done in the next quarter. Also in the next quarter the Lattice copyData and set functions will be extended to handle masks.
No particular problems. NFRA will continue to test AIPS++ on HP machines.
The core NRAO AIPS++ group is now Barnes (100%), Cornwell (100%), Garwood (100%), Holdaway(100%), Kemball(100%), Marson (100%), McMullin (100%), Schiebel (100%), Weatherall (50%), and Young (100%).
Peter:
Tim did the following:
Bob Garwood's primary responsibility is to oversee the single dish work which remains focussed on the dish environment as well as supporting the GBT.
Over the past 3 months (July through September) he has done the following work:
Brian Glendenning worked 75%+ for the MMA last quarter. His AIPS++ activities were modest: some image changes to support reporting ancillary information about the original observation, automated detection of object creation/deletion for the object GUI, some bug fixes, and preparation for the upcoming Beta release (which he will conduct).
Mark Holdaway joined the AIPS++ project in July and is improving the mosaicing capabilities in AIPS++. Currently, he is working on improving the voltage pattern classes. The classes will be general enough to handle 1-D radial beam profiles or 2-D beam images, correctly rotating the latter on the sky. In correctly treating the 2-D beams, we can correct for off-axis instrumental polarization and circular polarization beam squint. The classes will also permit solving for the primary beam shape via some form of self-calibration in the future.
Mark is also organizing a Mosaicing Techniques Meeting to be held in Socorro in July 1999.
Athol has become Deputy PM. He has taken responsibility for the oversight of the syntheis development, and he also manages the parallelization effort. His prime effort in coding was the development of the calibration object, calibrater. In addition, he continues to work on system-wide issues such as integration of applications.
Joseph McMullin's main responsibilities are managing the Green Bank installation of AIPS++ and supporting AIPS++-related GBT needs.
Developed a basic continuum commissioning software suite in AIPS++ as per the specifications of the McKinnon and Braatz memo on "Preliminary Single-Dish Continuum Requirements for AIPS++". The package reads AIPS++ table data and provides a set of functions from either the command line or a GUI interface. Available functions include: 1) system temperature measurement, 2) power spectrum calculation, 3) display of a tracking data, 4) measurement of antenna temperature, 4) analysis of focus observation, 5) analysis of tipping observation for opacity, and 6) analysis of a pointing observation. This package was presented at the GBT science workshop. A corresponding user's document was written for this software.
Wrote a script for reducing total power scans from the 140 ft/GBT spectral processor, and loading the reduced scans into a table readable by DISH for further analysis. This has been used for off-line reduction of observations at the 140 ft, as well as passband tests of the spectral processor.
Implemented a revised version of the AIPS++ RFI monitoring program based on comments from the RFI committee in GB. It's operation has been shown to the existing 140 ft/GBT operators and is in use.
Currently working on the remaining specifications in the McKinnon and Braatz memo, specifically implementation of tools for pointing analysis and modelling of the telescope structure.
Darrell Schiebel implemented all of the Glish changes and additions described in the section on progress in the last quarter.
In addition to all of the Glish changes, Darrell worked on a few other items. He wrote the GBT observer table parser which Rick Fisher drives with his GBT Observer interface. This Glish client parses the GBT observer table language which Rick developed, and returns Glish records which are arranged for easy use by Rick's scripts.
Following the departure of Jeff Uphoff, Darrell worked on a number of AIPS++ system/clean up items:
Kate was responsible for all changes made to the AIPS++ Web pages. She also wrote some documentation, and collaborated with Anantha on the format of the Getting Started document.
In this section, we give the names of people in the various AIPS++ groups and the nominal fraction of time allocated to AIPS++. These are as expected for the next quarter (and thus do not include the expected replacements for Briggs or Uphoff, or the new hires in Socorro).
The ATNF group is: Neil Killeen (75%, also local manager), Wim Brouw (20%), Mark Wieringa (30%) and David Barnes (75%).
The BIMA/NCSA group is: Harold Ravlin (10%), and Doug Roberts (100%). Another position is in the processing of being filled.
The NFRA group is: Ger van Diepen, Michael Haller, and Jan Noordam. The allocations are:
| Global AIPS++ | Local AIPS++ | Other | |
| Ger van Diepen: | 60% | 20% | 20% |
| Michael Haller: | 10% | 30% | 60% |
| Jan Noordam: | 30% | 40% | 30% |
The NRAO group is: Tim Cornwell (100%), Bob Garwood (100%), Mark Holdaway (100%), Athol Kemball (100%), Ralph Marson (100%), Joe McMullin (100%), Darrell Schiebel (100%), Kate Weatherall (50%) and Wes Young (100%). We have one open position in Green Bank and two new positions in Socorro (funded by NRL and NCSA).
Thus, in aggregate, we have 19 people contributing about 14.6 FTEs to the AIPS++ Project. Of these, 10 are employed by NRAO (one funded by NCSA), and contribute 9.5 FTEs. The numbers for the other partners are: ATNF 4 and 2, BIMA/NCSA 2 and 1.1, NFRA 3 and 1.9. There are four unfilled positions, one at BIMA/NCSA (the replacement for Briggs) and three at NRAO (GBT scientist, parallel applications developer, and an NRL-funded postdoc).
