NRAO AIPS++ Users Group Meeting Date: 2003-3-26 (Wednesday) Time: 1300 MST Video Hub: CV-conf (NOTE CHANGE: SOC, GB, TUC should call into CV-conf!) Rooms: SOC317/CV311/GB241/TUCN505 AIPS++ Threat Level is: Blue = guarded (cautious optimism) [downgraded from Red after Garching week, from Orange after Tech Review, and from Yellow after PDR] Agenda (full - we'll have to move through this!): 1. AIPS++ Technical Review (Steve, Joe, Tim) o The Interim Draft Report (3/17/2003) from the Technical Review Panel is available at: http://aips2.nrao.edu/projectoffice/aipstr/Interim.pdf They will provide the final report by April 1. o The results, which were summarized in my previous mail http://www.aoc.nrao.edu/~smyers/aips++/agenda/naug-20030312-agenda.txt are: - There is no "fundamental flaw" in the archictecture, and the toolkit is sound (though poorly integrated) - Focus should be changed from a general package toward fulfilling requirements of key future projects (ALMA,GBT,EVLA,etc.), including de-emphasis of support and outreach for current non-critical users - The development process should be modified, with use-case based requirements and testing, design reviews, auditing and acceptance, tied to the project milestones - The Project Management and Team should be strengthened, with Joe becoming full-time manager (sorry Joe) and adoption of standard management methodology (sorry Joe again), and more efficient use of the team personnel (less splitting of Consortium personell effort, filling of vacancies) - The Project Scientist role should be broadened and strengthened (sorry Steve), with a Consortium-level appointment equivalent to Project Manager, with responsibility for handling and prioritizing requirements and use cases, and with the authority for acceptance of deliverables and declaration of a release - Short-term (1-year) priorities should focus on significant improvements on reliability and stability, performance improvments, fulfillment of some key end-to-end VLA use cases (targeted for ALMA), with short-term de-emphasis of usability and interface improvements (pending technology upgrade, though studies should be carried out); cookbook development is seen to be important - Process with study of proposed technology change (based on ALMA ACS), with proof-of-concept for Fall 2003 and a design (followed by a review) in a year; this should not detract for other core work such as the above improvements or GBT support o The NAUG work of the past few years has played a key role in the success of this review, and we should plan to do even better in the future (see item 3 below). See the Interim Report for some nice recognition of all our work. o Note that as we had planned based on discussions with Fred last fall, the VLA testing is an important step toward ALMA. In fact, the panel was very clear on the need for VLA-based Use Cases (such as what Frazer had requested) for end-to-end testing. These are our marching orders for the next few months! Note that the audit we just did puts us in a good position to audit the use cases when we have them. 2. ALMA Computing PDR (Tim, Kumar, Steve) o Kumar and Lindsey presented the plans for ALMA Offline and Pipeline subsystems at the ALMA PDR last week in Tucson. Tim presented the Tech Review and NRAO response, and Sanjay the benchmarking results. These were well-received, and all indications are that we are on track. o The panel gave a verbal summary. The major AIPS++ related points were: - It's clear to them that AIPS++ has made a major course change in the last few months. - They didn't want to second-guess the TR report but they noted that the TR report said that AIPS++ should be very much driven by the needs of its clients. Thus they advised the ALMA Computing IPT that it should ensure that AIPS++ meets the real needs of ALMA (e.g. via the Offline subsystem scientist Debra) - They recommended an SSR audit of the capabilities of AIPS++ against the pipeline needs (similar to what I did for the Offline) - They welcomed the open architecture but suggested that we also might do well to work more to allow others to insert algorithms in AIPS++ (not clear what this meant) - They said that the project should state that it has adopted AIPS++ (this would be welcome!) o Debra (the ALMA SSR Offline Subsystem Scientist) is drawing up an ALMA Offline Testing Plan covering 2003-2006. This will rely upon NAUG pre-testing of things, and thus coordination is needed between our testing (below) and the ALMA testing. o There is common interest between ALMA and the NAUG in GUI look-and-feel, and we might want to think about coordination in thinking about this. 3. NAUG Plan 2003 (Steve) o I am in the process of drawing up a NAUG Plan for 2003, so that I can request from Jim & Greg part of peoples' time for NAUG activites. o I roughly need the equivalent of 8 people at 10%. A strawman plan asks for 2 @ 20%, 3 @ 10%, 2 @ 5%. o NAUG activities feature critically in the AIPS++ plans for the coming years, and I hope that as many of you as possible will agree to participate in the NAUG in a significant way this year (and next)! Note that these activities are important for NRAO, in particular for the key projects such as ALMA, GBT, and EVLA. o The current plan has 5 Subsystem Scientists in the areas: - Synthesis Calibration & Editing - Synthesis Imaging & Analysis - Single Dish Calibration & Editing - Single Dish Imaging & Analysis - General Interface & Documentation Note that the two pairs of Synthesis and Single Dish subsystems could also be broken down as Continuum and Spectral Line if we wanted them to follow the Use Cases more closely (discussion point). o The key deliverables for the NAUG this year are VLA-based Use Cases and testing based on them (targeted toward ALMA requirements), and GBT Use Cases, Requirements, and testing. I would like to get the first VLA (and GBT) use cases by June, and to have two periods of focused testing in May-July and Sep-Nov. o I plan to have a GUI Focus Group in 2003 to start thinking about storyboarding possible look and feel for integrated interfaces. Start thinking about how you would like GUIs to look for VLA and GBT! 4. VLA Requirements & Use Cases (Steve and Debra) o Thanks for your excellent work on this! o Latest versions incorporating contributions are in: http://www.aoc.nrao.edu/~smyers/aips++/vla/VLAoffline_audit.pdf http://www.aoc.nrao.edu/~smyers/aips++/vla/VLAoffline_audit.ps.gz http://www.aoc.nrao.edu/~smyers/aips++/vla/VLAoffline_audit.tex o In pretty good shape, I still need to go through and deal with incomplete items o In my latest draft above, I have make a stab at defining the modes of the VLA - this is included as Appendix A. Look this over and comment, as these will be needed for Use Cases. o I also included Frazer's wide-field requirements as Appendix B. Again, please look these over and think of how to turn these into Use Cases. o Use Cases: - these were requested by the Technical Review panel!!!! - it is hard in the current document to see whether a given path, (e.g. Frazer's wide-field case) is supported - should have a set of paths or use-cases written as a list of steps (which can be linked to requirements in the main document) which define a mode - put these as appendices to the requirements (like in the ALMA SW-11 document, see http://www.alma.nrao.edu/development/computing/docs/joint/0011/ssranduc.pdf - I will try to draft the first use case this coming month (e.g. continuum intensity), should have a few main mode use cases done by end of May and start testing 5. GBT Requirements & Use Cases (Steve and Joe) o The GBT SSR group requested (quite preciently!) that they tackle Use Cases for key modes instead of the broad requirements o The Use Cases are high-priority given the Technical Review! I would like to see the first set by end of May, with one case by end of April. o I propose that Jay Lockman coordinates the Use Case collection at GB Volunteers needed to define modes and draft Use Cases. o "NAUG" GBT requirements/evaluation draft progress report (Steve) - draft was produced on 10 Jan. Can find this at: http://www.aoc.nrao.edu/~smyers/aips++/gbt/naug_gbt_eval.tex http://www.aoc.nrao.edu/~smyers/aips++/gbt/naug_gbt_eval.pdf http://www.aoc.nrao.edu/~smyers/aips++/gbt/naug_gbt_eval.ps.gz These reqs were drafted by me and are rough and incomplete. In particular the priorities and timescales were somewhat arbitrary. These are under comment by the GB folks (Jay Lockman). Note: Harvey's old SDcalc email is incorporated as Appendix A. NAUG comments welcome also! 6. AIPS++ news o The AIPS++ Project Office prototype is online, see http://projectoffice.aips2.nrao.edu Available at the Project Office: - Project Book - Technical Review materials - Stable Release Calendar and Development Plans - Current Stable snapshot - Change logs - other stuff! Comments welcome... o The current stable snapshot is v1.8 Build 569 2003/03/22, see http://aips2.nrao.edu/daily/docs/reference/updates.html o Benchmarking: Sanjay wrote up a description of the benchmarking for the AIPS++ Technical Review and the ALMA PDR. See his presentation at: http://aips2.nrao.edu/projectoffice/aipstr/aipstr_bench_files/v3_document.htm Also see Section 6.3 of the Technical Review Documentation. Sanjay has made excellent progress on identification and fixing of problem areas in the code, and in imager in particular. There does not appear to be a "fundamental flaw" in the AIPS++ architecture, and we should be able to bring the performance to the level of AIPS through the careful process that Sanjay has started. 7. Upcoming meetings and deadlines: o Apr 8-16 Joe, Steve & Tim to CV/GB o Apr 14-16 NRAO Visiting Committee (GB) o Jun 13-14 NRAO Users Committee (SOC) The agendas for past NAUG meetings are archived at: http://www.aoc.nrao.edu/~smyers/aips++/agenda/ The minutes for past NAUG meetings are archived at: http://www.aoc.nrao.edu/~smyers/aips++/minutes/